|
Journal Publications
|
| 1 |
|
Ioannis Tsoulos, Dimitris Gavrilis, Euripidis Glavas, "Neural Network Construction and Training using Grammatical Evolution", Neurocomputing, 2008  Abstract:  |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
|
| 2 |
|
Dimitris Gavrilis, Ioannis G. Tsoulos, Evangelos Dermatas, "Selecting and Constructing Features Using Grammatical Evolution", Pattern Recognition Letters, 2008 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
|
| 3 |
|
Ioannis G. Tsoulos, Dimitris Gavrilis, Evangelos Dermatas, "GDF v2.0, an enhanced version of GDF", Computer Physics Communications, Issue 177, 2007, pp. 976-977 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
|
| 4 |
|
George Georgoulas, Dimitris Gavrilis, Ioannis G. Tsoulos, Chrysostomos Stylios, Joao Bernardes, Peter P. Groumpos, "Novel approach for fetal heart rate classification introducing grammatical evolution", Biomedical Signal Processing and Control, Issue 2, 2007, pp. 69-79 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
|
| 5 |
|
Ioannis Tsoulos, Dimitris Gavrilis, Evangelos Dermatas, "GDF: A tool for function estimation through grammatical evolution", Computer Physics Communication, Vol. 174, Issue 7, 2006, pp. 555-559 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
|
| 6 |
|
Dimitris Gavrilis, Evangelos Dermatas, "Real-time detection of distributed denial-of-service attacks using RBF networks and statistical features", Computer Networks, Vol. 48, 2005, pp. 235-245 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
|
|
Conference Publications
|
| 1 |
|
Dimitris Gavrilis, Ioannis G. Tsoulos, Evangelos Dermatas, "Feature Selection for Robust Detection of Distributed Denial-of-Service Attacks Using Genetic Algorithms", Lecture Notes in Computer Science, Vol. 3025, 276-281, pp. 2004 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
|
| 2 |
|
Dimitris Gavrilis, Constantia Kakali, Christos Papatheodorou, "Enhancing Library Services with Web 2.0 functionalities ", European Conference on Research and Advanced Technology for Digital Libraries (ECDL 2008), 2008 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
|
| 3 |
|
Antonios Deligiannakis, Dimitris Gavrilis, Yannis Kotidis, Christos Papatheodorou, Stavros Angelis, Panos Constantopoulos, Ion Androutsopoulos, Costis Dallas, "DCC&U: An Extended Digital Curation Lifecycle Model", 4th International Digital Curation Conference , 2008 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
|
| 4 |
|
Dimitris Gavrilis, Giannis Tsakonas, Christos Papatheodorou, "Designing Interoperable Museum Information Systems", International Conference on Virtual Systems and MultiMedia (VSMM), 2008 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
|
| 5 |
|
Panos Constantopoulos, Costis Dallas, Petern Doorn, Dimitris Gavrilis, Andreas Gros, Georgios Stylianou, "PREPARING DARIAH", International Conference on Virtual Systems and MultiMedia (VSMM), 2008 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
|
| 6 |
|
Spyros Veronikis, Dimitris Gavrilis, Kyriaki Zoutsou, Christos Papatheodorou, "Using Handhelds to Search in Physical and Digital Information Spaces", The Second International Conference on Mobile Ubiquitous Computing, Systems, Services and Technologies (UBICOMM 2008), 2008 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
|
| 7 |
|
Dimitris Gavrilis, Ioannis Chatzis and Evangelos Dermatas, "Flash Crowd Detection Using Decoy Hyperlinks", 2007 IEEE International Conference On Networking, Sensing and Control - ICNSC07, 2007 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
|
| 8 |
|
Dimitris Gavrilis, Evangelos Dermatas, "Neural Recognition and Genetic Features Selection for Robust detection of E-mail Spam", Lecture Notes in Computer Science, 2006 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
|
| 9 |
|
Dimitris Gavrilis, Ioannis S. Chatzis, Evangelos Dermatas, "Detection of Web Denial-of-Service Attacks using decoy hyperlinks", 5th International Symposium on Communication Systems, Networks and Digital Signal Processing (CSNDSP), 2006 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
|
| 10 |
|
Ioannis Tsoulos, Georgoulas Georgoulas, Dimitris Gavrilis, Chrysostomos Stylios, Joao Bernardes, Peter Groumpos, "Introducing Grammatical Evolution for FHR analysis and classification", 3rd IEEE Conference on Intelligent Systems (IEEE-IS 06), 2006 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
| Electronic fetal monitoring is an essential tool for fetal surveillance during labour. It is mainly based on the monitoring and evaluation of the Fetal Heart Rate, (FHR) which is a biosignal that has to be interpreted on line. Evaluation and interpretation of FHR gives an indication of the fetus health status. A lot of research efforts have been done towards the development of automatic and reliable methods for processing and evaluating FHR. This research work, introduces an integrated methodology for processing and classifying FHR based on the novel approach of grammatical evolution for feature construction and selection. The proposed methodology is presented, which then is applied to data set. Experimental results are promising paving the way for further research in that direction. |
|
| 11 |
|
Ioannis Chatzis, Dimitris Gavrilis, Evangelos Dermatas, "Spectral Characterization of Digital Cameras using Genetic Algorithms", 2nd I*PROMS Virtual International Conference on Intelligent Production Machines and Systems, 2006 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
| Electronic fetal monitoring is an essential tool for fetal surveillance during labour. It is mainly based on the monitoring and evaluation of the Fetal Heart Rate, (FHR) which is a biosignal that has to be interpreted on line. Evaluation and interpretation of FHR gives an indication of the fetus health status. A lot of research efforts have been done towards the development of automatic and reliable methods for processing and evaluating FHR. This research work, introduces an integrated methodology for processing and classifying FHR based on the novel approach of grammatical evolution for feature construction and selection. The proposed methodology is presented, which then is applied to data set. Experimental results are promising paving the way for further research in that direction. |
| In camera characterization a number of techniques is applied to minimize the impact of different hardware and software implementation in image acquisition systems, and preserve colour distortion between devices. In this paper, a new method for estimating spectral response from illuminated images is presented and evaluated based on genetic algorithms. The optimization criterion minimizes the maximum difference between a mixture of Gaussian functions and the real spectral response. A genetic optimization process estimates the parameters of the gaussian mixture, implemented in Java using the tournament selection method. The experimental results show significant improvement of the proposed spectral estimation method over the well-known PCA method using the first six to ten most significant eigenvectors in the presence of three additive noises: the noise which is statistically independent to the intensity level, the signal dependent noise, and the digitization noise. |
|
| 12 |
|
Dimitris Gavrilis, Ioannis Tsoulos, Evangelos Dermatas, "Stochastic Classification of Scientific Abstracts", 11-th International Conference Speech and Computer (SPECOM 2006), 2006 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
| Electronic fetal monitoring is an essential tool for fetal surveillance during labour. It is mainly based on the monitoring and evaluation of the Fetal Heart Rate, (FHR) which is a biosignal that has to be interpreted on line. Evaluation and interpretation of FHR gives an indication of the fetus health status. A lot of research efforts have been done towards the development of automatic and reliable methods for processing and evaluating FHR. This research work, introduces an integrated methodology for processing and classifying FHR based on the novel approach of grammatical evolution for feature construction and selection. The proposed methodology is presented, which then is applied to data set. Experimental results are promising paving the way for further research in that direction. |
| In camera characterization a number of techniques is applied to minimize the impact of different hardware and software implementation in image acquisition systems, and preserve colour distortion between devices. In this paper, a new method for estimating spectral response from illuminated images is presented and evaluated based on genetic algorithms. The optimization criterion minimizes the maximum difference between a mixture of Gaussian functions and the real spectral response. A genetic optimization process estimates the parameters of the gaussian mixture, implemented in Java using the tournament selection method. The experimental results show significant improvement of the proposed spectral estimation method over the well-known PCA method using the first six to ten most significant eigenvectors in the presence of three additive noises: the noise which is statistically independent to the intensity level, the signal dependent noise, and the digitization noise. |
| In this paper a method for feature selection and scientific documents classification is presented. The initial selection of features is performed by measuring their entropy and the final selection is implemented using a genetic algorithm. A neural classifier is used to ensure robust classification rate even in case of complex cluster structure. For the evaluation of the proposed method, 650 abstracts are derived from the PUBMED and were used for features selection and classification into 5 categories. A classification rate of 85% is achieved using only 20 features. |
|
| 13 |
|
Dimitris Gavrilis, Ioannis Tsoulos, Evangelos Dermatas, "Evolutionary Grammar Induction for Protein Relation Extraction", 11-th International Conference Speech and Computer (SPECOM 2006), 2006 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
| Electronic fetal monitoring is an essential tool for fetal surveillance during labour. It is mainly based on the monitoring and evaluation of the Fetal Heart Rate, (FHR) which is a biosignal that has to be interpreted on line. Evaluation and interpretation of FHR gives an indication of the fetus health status. A lot of research efforts have been done towards the development of automatic and reliable methods for processing and evaluating FHR. This research work, introduces an integrated methodology for processing and classifying FHR based on the novel approach of grammatical evolution for feature construction and selection. The proposed methodology is presented, which then is applied to data set. Experimental results are promising paving the way for further research in that direction. |
| In camera characterization a number of techniques is applied to minimize the impact of different hardware and software implementation in image acquisition systems, and preserve colour distortion between devices. In this paper, a new method for estimating spectral response from illuminated images is presented and evaluated based on genetic algorithms. The optimization criterion minimizes the maximum difference between a mixture of Gaussian functions and the real spectral response. A genetic optimization process estimates the parameters of the gaussian mixture, implemented in Java using the tournament selection method. The experimental results show significant improvement of the proposed spectral estimation method over the well-known PCA method using the first six to ten most significant eigenvectors in the presence of three additive noises: the noise which is statistically independent to the intensity level, the signal dependent noise, and the digitization noise. |
| In this paper a method for feature selection and scientific documents classification is presented. The initial selection of features is performed by measuring their entropy and the final selection is implemented using a genetic algorithm. A neural classifier is used to ensure robust classification rate even in case of complex cluster structure. For the evaluation of the proposed method, 650 abstracts are derived from the PUBMED and were used for features selection and classification into 5 categories. A classification rate of 85% is achieved using only 20 features. |
| A novel method is presented for protein relation extraction from scientific abstracts. The proposed method is based on Meta-Grammars, a novel method for grammar inference that uses genetic programming and a BNF description to discover a tree representation of sentence structure that can be used for information extraction. A series if transformations are applied to the original corpus before the Meta-Grammars genetic algorithm is applied. The proposed method is evaluated against extracting protein relations from scientific abstracts and it is shown that it requires a train corpus which has minimum requirements from field experts and giving precision of 79.165%. |
|
| 14 |
|
Ioannis Tsoulos, Dimitris Gavrilis, Euripidis Glavas, "Neural Network Construction using Grammatical Evolution", 5th IEEE International Symposium on Signal Processing and Information Technology, 2005 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
| Electronic fetal monitoring is an essential tool for fetal surveillance during labour. It is mainly based on the monitoring and evaluation of the Fetal Heart Rate, (FHR) which is a biosignal that has to be interpreted on line. Evaluation and interpretation of FHR gives an indication of the fetus health status. A lot of research efforts have been done towards the development of automatic and reliable methods for processing and evaluating FHR. This research work, introduces an integrated methodology for processing and classifying FHR based on the novel approach of grammatical evolution for feature construction and selection. The proposed methodology is presented, which then is applied to data set. Experimental results are promising paving the way for further research in that direction. |
| In camera characterization a number of techniques is applied to minimize the impact of different hardware and software implementation in image acquisition systems, and preserve colour distortion between devices. In this paper, a new method for estimating spectral response from illuminated images is presented and evaluated based on genetic algorithms. The optimization criterion minimizes the maximum difference between a mixture of Gaussian functions and the real spectral response. A genetic optimization process estimates the parameters of the gaussian mixture, implemented in Java using the tournament selection method. The experimental results show significant improvement of the proposed spectral estimation method over the well-known PCA method using the first six to ten most significant eigenvectors in the presence of three additive noises: the noise which is statistically independent to the intensity level, the signal dependent noise, and the digitization noise. |
| In this paper a method for feature selection and scientific documents classification is presented. The initial selection of features is performed by measuring their entropy and the final selection is implemented using a genetic algorithm. A neural classifier is used to ensure robust classification rate even in case of complex cluster structure. For the evaluation of the proposed method, 650 abstracts are derived from the PUBMED and were used for features selection and classification into 5 categories. A classification rate of 85% is achieved using only 20 features. |
| A novel method is presented for protein relation extraction from scientific abstracts. The proposed method is based on Meta-Grammars, a novel method for grammar inference that uses genetic programming and a BNF description to discover a tree representation of sentence structure that can be used for information extraction. A series if transformations are applied to the original corpus before the Meta-Grammars genetic algorithm is applied. The proposed method is evaluated against extracting protein relations from scientific abstracts and it is shown that it requires a train corpus which has minimum requirements from field experts and giving precision of 79.165%. |
| A method which is based on grammatical evolution is presented in this paper for the construction of artificial neural networks (ANNs). The method is capable to construct ANNs with an arbitrary number of hidden levels or even recurrent neural networks. The efficiency of the method is tested on a series of classification and regression problems and the results are compared against traditional neural networks. |
|
| 15 |
|
Dimitris Gavrilis, Ioannis Tsoulos, George Georgoulas, Euripidis Glavas, "Classification of Fetal Heart Rate using Grammatical Evolution", IEEE 2005 Workshop on Signal Processing Systems (SiPS 2005), 2005 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
| Electronic fetal monitoring is an essential tool for fetal surveillance during labour. It is mainly based on the monitoring and evaluation of the Fetal Heart Rate, (FHR) which is a biosignal that has to be interpreted on line. Evaluation and interpretation of FHR gives an indication of the fetus health status. A lot of research efforts have been done towards the development of automatic and reliable methods for processing and evaluating FHR. This research work, introduces an integrated methodology for processing and classifying FHR based on the novel approach of grammatical evolution for feature construction and selection. The proposed methodology is presented, which then is applied to data set. Experimental results are promising paving the way for further research in that direction. |
| In camera characterization a number of techniques is applied to minimize the impact of different hardware and software implementation in image acquisition systems, and preserve colour distortion between devices. In this paper, a new method for estimating spectral response from illuminated images is presented and evaluated based on genetic algorithms. The optimization criterion minimizes the maximum difference between a mixture of Gaussian functions and the real spectral response. A genetic optimization process estimates the parameters of the gaussian mixture, implemented in Java using the tournament selection method. The experimental results show significant improvement of the proposed spectral estimation method over the well-known PCA method using the first six to ten most significant eigenvectors in the presence of three additive noises: the noise which is statistically independent to the intensity level, the signal dependent noise, and the digitization noise. |
| In this paper a method for feature selection and scientific documents classification is presented. The initial selection of features is performed by measuring their entropy and the final selection is implemented using a genetic algorithm. A neural classifier is used to ensure robust classification rate even in case of complex cluster structure. For the evaluation of the proposed method, 650 abstracts are derived from the PUBMED and were used for features selection and classification into 5 categories. A classification rate of 85% is achieved using only 20 features. |
| A novel method is presented for protein relation extraction from scientific abstracts. The proposed method is based on Meta-Grammars, a novel method for grammar inference that uses genetic programming and a BNF description to discover a tree representation of sentence structure that can be used for information extraction. A series if transformations are applied to the original corpus before the Meta-Grammars genetic algorithm is applied. The proposed method is evaluated against extracting protein relations from scientific abstracts and it is shown that it requires a train corpus which has minimum requirements from field experts and giving precision of 79.165%. |
| A method which is based on grammatical evolution is presented in this paper for the construction of artificial neural networks (ANNs). The method is capable to construct ANNs with an arbitrary number of hidden levels or even recurrent neural networks. The efficiency of the method is tested on a series of classification and regression problems and the results are compared against traditional neural networks. |
| Abstract—There is an ongoing effort to develop advanced methods and computer-based systems to assist obstetricians in the difficult task of feature extraction and classification of the Cardiotocogram (CTG), which is the most widely used Electronic Fetal Monitoring (EFM) method worldwide. A novel method for feature construction is presented for efficient classification of CTG based on information extracted from Fetal Heart Rate (FHR) signal. The proposed method is based on grammatical evolution in order to construct new features from existing ones using nonlinear transformations. This method is tested on a data set of intrapartum cases achieving accuracy of 92.5%. |
|
| 16 |
|
Dimitris Gavrilis, Evangelos Dermatas, "Part-of-Speech Tagging in Molecular Biology Scientific Abstracts Using Morphological and Contextual Statistical Information", Lecture Notes in Computer Science, Vol. 3025, 2004, pp. 371-380 Abstract: |
| |
|
| The term neural network evolution usually refers to network topology evolution leaving the network’s parameters to be trained using conventional algorithms. In this paper we present a new method for neural network evolution that evolves the network topology along with the network parameters. The proposed method uses grammatical evolution to encode both the network and the parameters space. This allows for a better description of the network using a formal grammar allowing the network architect to shape the resulting search space in order to meet each problem requirements. The proposed method is compared to other three methods for neural network training and is evaluated using 9 known classification problems and 9 known regression problems. In all 18 datasets the proposed method outperforms its competitors. |
| A novel method for feature selection and construction is introduced. The method improves the classification accuracy, utilizing the well-established technique of Grammatical Evolution by creating non-linear mappings of the original features to artificial ones in order to improve the effectiveness of artificial intelligence tools such as Multi-Layer Perceptron (MLP), Radial-Basis Function (RBF) neural networks and Nearest Neighbor (KNN) classifier. The proposed method has been applied on a series of classification and regression problems and an experimental comparison is carried out against the accuracy obtained on the original features as well as on features created by the PCA method. |
| An improved version of the function estimation program GDF is presented. The main enhancements of the new version include: multi-output function estimation, capability of defining custom functions in the grammar and selection of the error function. The new version has been evaluated on a series of classification and regression datasets, that are widely used for the evaluation of such methods. It is compared to two known neural networks and outperforms them in 5 (out of 10) datasets. |
| Fetal heart rate (FHR) variations reflect the level of oxygenation and blood pressure of the fetus. Electronic Fetal Monitoring (EFM), the continuous monitoring of the FHR, was introduced into clinical practice in the late 1960s and since then it has been considered as an indispensable tool for fetal surveillance. However, EFM evaluation and its merit is still an open field of controversy, mainly because it is not consistently reproducible and effective. In this work, we present a novel method based on grammatical evolution to discriminate acidemic from normal fetuses, utilizing features extracted from the FHR signal during the minutes immediately preceding delivery. The proposed method identifies linear and nonlinear correlations among the originally extracted features and creates/constructs a set of new ones, which, in turn, feed a nonlinear classifier. The classifier, which also uses a hybrid method for training, along with the constructed features was tested using a set of real data achieving an overall performance of 90% (specificity = sensitivity = 90%). |
| This article introduces a tool for data fitting that is based on genetic programming and especially on the grammatical evolution technique. The user needs to input a series of points and the accompanied dimensionality n and the tool will produce via the genetic programming paradigm a function f :Rn →R which is an approximate solution to the symbolic regression problem. The tool is entirely written in ANSI C++ and it can be installed in any UNIX system. |
| In this paper we present and evaluate a Radial-basis-fucntion neural network detector for Distributed-Denial-of-Service (DDoS) attacks in public networks based on statistical features estimated in short-time window analysis of the incoming data packets. A small number of generalized descriptors were used to describe the statistical behaviour of DDoS attacts, and an accurate classification is achieved using the Radial-basis-function neural networks (RBF-NN). The proposed method is evaluated in a simulated public network and showed detection rate better than 98% of DDoS attacs using only three statistical features estimated from one window of data packets of 1 second length. The same type of experiments were carried out on a real network giving significantly better results: a 100% DDoS detection rate is achieved followed by a 0% of false alarm rate using different statistical descriptors and training conditions for the RBF-NN. |
| In this paper we present a robust neural network detector for Distributed Denial-of-Service (DDoS) attacks in computers providing Internet services. A genetic algorithm is used to select a small number of efficient features from an extended set of 44 statistical features, which are estimated only from the packet headers. The genetic evaluation produces an error-free neural network DDoS detector using only 14 features. Moreover, the experimental results showed that the features that best qualify for DDoS detection are the SYN and URG flags, the probability of distinct Source Ports in each timeframe, the number of packets that use certain port ranges the TTL and the window size in each timeframe. |
| In this paper, a prototype of an Online Public Access Catalog (OPAC) is presented. This new OPAC features new functionalities and utilizes web 2.0 technologies in order to deliver improved search and retrieval services. Some of these new services include social tag annotations, user opinions and ranks and tag-based similarity searches. The prototype is evaluated by a user group through questionnaires, interviews and with the system's integrated logging mechanism. The results are encouraging enough and show that Library 2.0 technologies seem to be acceptable by the majority of the users. |
| The proliferation of web, database and electronic-socialising technologies has enabled us to produce, publish and exchange digital assets at an enormous rate. This vast amount of information that is either digitized or born-digital needs to be collected, organized and preserved in a way that ensures that our digital assets and the information that they carry remain available for future use. Digital curation has emerged as a new inter-disciplinary practice that seeks to set guidelines for disciplined management of information. In this paper we review two recent models for digital curation introduced by the Digital Curation Centre (DCC) of UK and the Digital Curation Unit (DCU) of the Athena Research Centre. We then propose a fusion of the two models that highlights the need to extend the digital curation life-cycle by: (a) provisions for the registration of usage experience, (b) an action of knowledge-enhancement, and (c) controlled vocabularies used by convention to denote concepts, properties and relations. The objective of the proposed extensions is twofold: Firstly, to upgrade and, mainly, to homogenize the research efforts and approaches on the digital curation domain. Secondly, to provide a basis for a wide and fruitful discussion on the development of a common research agenda as well as for an international cooperation and coordination of the research communities. |
| Museum collections are characterized by heterogeneity, since they usually host a plethora of objects of categories, while each of them requires different description policies and metadata standards. Moreover the museum records, which keep the history and evolution of the hosted collections, request proactive curation in order to preserve this rich and diverse information. In this paper, the architecture of an innovative museum information system, as well as its implementation details is presented. In particular the requirements and the system architecture are presented along with the problems that were encountered. The main directions of the system design are (a) to increase interoperability levels and therefore assist proactive curation and (b) to enhance navigation by the usage of handheld devices. The first direction is satisfied by the design of a rich metadata schema based on the CIDOC/CRM standard. The second direction is fulfilled by the implementation of a module, which integrates the museum database with a subsystem appropriate to support user navigation into the museum floors and rooms. The module is expressed as a navigation functionality, which is accessed through handheld devices and peripherals, such as PDAs and RFID tags. The proposed system is functional and operates into the Solomos Museum, situated in Zakynthos island, Greece. |
| In this paper, a preparatory project for an integrated European research infrastructure in the humanities is presented. This project, Preparing for the construction of the Digital Research Infrastructure for the Arts and Humanities - or Preparing DARIAH for short, is part of the ESFRI e-infrastructures programme and supports the emergence of a new collaborative framework in which researchers are able to maximise the impact of their work on the international stage and aims at providing the foundations for the timely construction of the infrastructure requisite for the arts, humanities and cultural heritage communities in the digital age. DARIAH uses an interdisciplinary approach and involves tackling a number of interrelated issues such as strategic, organisational, financial, technical and conceptual in order to facilitate long-term access to and use of all European humanities and cultural heritage information for the purposes of enhancing and expanding research, thereby increasing our knowledge and understanding of our histories, heritage, languages and cultures. The DARIAH network will act as a place where the incubation of new ideas and ways of working can be facilitated and developed, and then transitioned into established organisations thus ensuring long-term sustainability and stability and the integration of these methods and techniques into everyday research practice. DARIAH will support research practitioners at all stages in the research process, and at differing levels of sophistication, from beginners through to those employing advanced techniques and methodologies. |
| In recent years a wealth of information is becoming available thanks to computer and networking technology. Modern libraries incorporate in their collections information content in both physical and digital form. Meanwhile, mobile computing enables the library patrons to access that content anytime, anywhere. In this paper we present the design procedure of a new library service that supports users in seeking information in hybrid collections while being in the stacks, thus enabling content retrieval from a unified information space. Moreover an evaluation model and methodology and the results of an experimental procedure are presented aiming to assess the user satisfaction for the new service. |
| In this paper a novel method for detecting Denial of Service attacks (DoS) on web services are presented and evaluated by using decoy hyperlinks embedded in web pages. The Decoys are hyperlinks without semantic information or are invisible to the human user, acting like traps for DoS attacks because a human user would never follow them. An attack on a web server is detected when such hyperlink is followed. This approach has significant advantages over other approaches like graphic Turing tests, it is transparent to the user, it can be used on general-purpose web sites and retains the web site’s usability. The proposed method has been evaluated using both real and simulated web sites and the results show false positive rates that are less than 10-4. The aspects of this new method are discussed and some experimental results are presented. |
| In this paper a method for feature selection and classification of email spam messages is presented. The selection of features is performed in two steps: The selection is performed by measuring their entropy and a fine-tuning selection is implemented using a genetic algorithm. In the classification process, a Radial Basis Function Network is used to ensure robust classification rate even in case of complex cluster structure. The proposed method shows that, when using a two-level feature selection, a better accuracy is achieved than using one-stage selection. Also, the use of a lemmatizer or a stop-word list gives minimal classification improvement. The proposed method achieves 96-97% average accuracy when using only 20 features out of 15000. |
| Abstract-In this paper a method for detecting Denial-of-Service attacks in Web sites is presented. The detection of Web attacks are distinguished from normal user patterns by inserting decoy hyperlinks into some key pages in the Website. Typical types of decoy hyperlinks are described and experimental results derived from real Web-sites gives the extremely low false positive rate of 0.0421%. A method for selecting an effective and minimal number of decoy hyperlinks and pages is also presented and evaluated in real and simulated data. |
| Electronic fetal monitoring is an essential tool for fetal surveillance during labour. It is mainly based on the monitoring and evaluation of the Fetal Heart Rate, (FHR) which is a biosignal that has to be interpreted on line. Evaluation and interpretation of FHR gives an indication of the fetus health status. A lot of research efforts have been done towards the development of automatic and reliable methods for processing and evaluating FHR. This research work, introduces an integrated methodology for processing and classifying FHR based on the novel approach of grammatical evolution for feature construction and selection. The proposed methodology is presented, which then is applied to data set. Experimental results are promising paving the way for further research in that direction. |
| In camera characterization a number of techniques is applied to minimize the impact of different hardware and software implementation in image acquisition systems, and preserve colour distortion between devices. In this paper, a new method for estimating spectral response from illuminated images is presented and evaluated based on genetic algorithms. The optimization criterion minimizes the maximum difference between a mixture of Gaussian functions and the real spectral response. A genetic optimization process estimates the parameters of the gaussian mixture, implemented in Java using the tournament selection method. The experimental results show significant improvement of the proposed spectral estimation method over the well-known PCA method using the first six to ten most significant eigenvectors in the presence of three additive noises: the noise which is statistically independent to the intensity level, the signal dependent noise, and the digitization noise. |
| In this paper a method for feature selection and scientific documents classification is presented. The initial selection of features is performed by measuring their entropy and the final selection is implemented using a genetic algorithm. A neural classifier is used to ensure robust classification rate even in case of complex cluster structure. For the evaluation of the proposed method, 650 abstracts are derived from the PUBMED and were used for features selection and classification into 5 categories. A classification rate of 85% is achieved using only 20 features. |
| A novel method is presented for protein relation extraction from scientific abstracts. The proposed method is based on Meta-Grammars, a novel method for grammar inference that uses genetic programming and a BNF description to discover a tree representation of sentence structure that can be used for information extraction. A series if transformations are applied to the original corpus before the Meta-Grammars genetic algorithm is applied. The proposed method is evaluated against extracting protein relations from scientific abstracts and it is shown that it requires a train corpus which has minimum requirements from field experts and giving precision of 79.165%. |
| A method which is based on grammatical evolution is presented in this paper for the construction of artificial neural networks (ANNs). The method is capable to construct ANNs with an arbitrary number of hidden levels or even recurrent neural networks. The efficiency of the method is tested on a series of classification and regression problems and the results are compared against traditional neural networks. |
| Abstract—There is an ongoing effort to develop advanced methods and computer-based systems to assist obstetricians in the difficult task of feature extraction and classification of the Cardiotocogram (CTG), which is the most widely used Electronic Fetal Monitoring (EFM) method worldwide. A novel method for feature construction is presented for efficient classification of CTG based on information extracted from Fetal Heart Rate (FHR) signal. The proposed method is based on grammatical evolution in order to construct new features from existing ones using nonlinear transformations. This method is tested on a data set of intrapartum cases achieving accuracy of 92.5%. |
| In this paper a probabilistic tagger for molecular biology related abstracts is presented and evaluated. The system consists of three modules: a rule based molecular-biology names detector, an unknown words handler, and a Hidden Markov model based tagger which are used to annotate the corpus with an extended set of grammatical and molecular biology tags. The complete system has been evaluated using 500 randomly selected abstracts from the MEDLINE database. The F-score for the molecular-biology names detector was 0.95, and the annotation rate was greater than 93% in all experiments using the Viterbi algorithm. The best annotation rate of 97.34% is achieved using a Pubmed dictionary. |
|
.png?1254156785 "User Profile Photo")