Search

This paper introduces ELEON, an editor that allows the enrichment of OWL ontologies with linguistic and user-related annotations. The enriched ontologies are used by natural language generation (NLG) engines to generate textual descriptions of the objects represented in the ontologies in the selected language and according to user's model. ELEON provides a well-defined interface that can be used by different NLG engines. The paper presents the relevant functionalities of ELEON, describes the provided interface to NLG engines and discusses the advantages of exploiting such enriched ontologies in NLG.

This paper develops the idea that variable ordering heuristics can be classified on the basis of a small number of distinguishable actions, and that while specific heuristics may be classified differently depending on the problem type, the basic actions that determine their classification are the same. Previous work demonstrated two basic categories of heuristics, and that problems in an apparently homogeneous problem set differ in their amenability to heuristics of different types. The present paper shows that these heuristic actions, which may be described as building up contention and propagating effects, have distinct values for descriptive measures such as depth of failure and the depth at which a problem becomes tractable, that reflect differences in the rapidity of their effects with respect to search depth. Heuristics behave similarly with respect to their basic actions across a wide range of propagation, from simple backtracking to maintained arc consistency. The propagation- of-effects type of action is closely related to the "simplification hypothesis" of Hooker and Vinay. This work contributes to the goals of explaining heuristic performance and putting heuristic selection on a rational basis.

Most research on preference elicitation, preference reasoning and preference query languages design focus mainly on preferences over single objects represented by relational tuples. An increasing interest on preferences over more complex structures like sets of objects has arised in recent papers. However, most recent applications deal with more sophisticated complex structured objects, like sequences, trees and graphs. In this paper, we introduce TPref, a formalism to reason with qualitative conditional preferences over sequences of objects indexed in time. TPref generalizes the CP-Nets formalism by allowing temporal conditional preferences besides the static rules used in CP-Nets. An algorithm for computing optimal sequences of objects satisfying a set of temporal constraints is also presented.

This paper presents a new algorithm for binarization of historical document images. It is adjusted to deal with the particular case of documents which are written in both sides of the paper. The digitization of these documents in general brings to the digital image this back-to-front interference. In order to deal with it, a thresholding algorithm is proposed based on color quantization by genetic algorithms and image fidelity analysis. The method achieved better results than other well-known algorithms.

In this paper we present a new incremental algorithm for building neighborhood graphs between data. It is inspired from the self-assembly behavior observed in real ants where ants progressively become attached to an existing support and then successively to other attached ants. Each artificial ant represents one data. The way ants move and build a graph depends on the similarity between the data. We have compared our results to those obtained by the relative neighborhood algorithm on several databases (either artificial or real), and we show that our method is competitive especially with respect to execution times.

The software cost estimation process is one of the most critical managerial activities related to project planning, resource allocation and control. As software development is a highly dynamic procedure, the difficulty of providing accurate cost estimations tends to increase with development complexity. The inherent problems of the estimation process stem from its dependence on several complex variables, whose values are often imprecise, unknown, or incomplete, and their interrelationships are not easy to comprehend. Current software cost estimation models do not inspire enough confidence and accuracy with their predictions. This is mainly due to the models' sensitivity to project data values, and this problem is amplified because of the vast variances found in historical project attribute data. This paper aspires to provide a framework for evolving value ranges for cost attributes and attaining mean effort values using the Al-oriented problem-solving approach of genetic algorithms, with a twofold aim. Firstly, to provide effort estimations by analogy to the projects classified in the evolved ranges and secondly, to identify any present correlations between effort and cost attributes.

Snapshot data have become ubiquitous, e.g., maps, images and videos. By extracting interesting features from snapshot data and analyzing their relative orientations and proximities, we can discover important structure configuration information among groups of features in a snapshot database. In this paper, we introduce a class of pattern called orientation-sensitive patterns, which occur in many applications ranging from weather study, sport game analysis to medical image processing. We examine three approaches to discover orientation-sensitive patterns. We show that the first apriori-based approach is expensive while the second enumeration-based approach is memory intensive. The third approach decomposes an orientation- sensitive pattern into an H-list and a V-list, which greatly simplifies the mining process. Extensive experiment studies show that the third method is more efficient and scalable than the apriori and enumeration algorithms. We also present case studies on soccer game snapshots to demonstrate the interesting patterns discovered.

Graph classification is an important data mining task that has attracted considerable attention recently. This paper presents a probabilistic substructure-based approach for classifying graph-based data. More specifically, we use a frequent subgraph mining algorithm to extract substructure based descriptors and apply the maximum entropy principle to build a classification model from the frequent subgraphs. We perform extensive experiments to compare the performance of the proposed approach against existing feature vector methods using AdaBoost and support vector machine.

Path relinking is an evolutionary approach for solving a diverse array of optimization problems. The NP-complete Steiner tree problem is a fundamental component of many real world telecommunication applications and is a critical structure underlying multicast communications. This paper describes a novel path relinking algorithm for the delay- constrained least-cost multicast routing problem. Computational results on various random generated networks indicate that our proposed path relinking algorithm results in the best overall performance with regard to total tree cost in comparison with other existing algorithms and heuristics.

One important and challenging problem in human interaction is the communication between blind and deaf individuals. The challenge here involves several cases: (i) first case is a deaf person usually does not speak in order a blind person to hear him/her; (ii) second case is when a blind person speaks a deaf person cannot hear; (iii) third case is when a deaf person makes sign language signs a blind person cannot see them. Thus, this paper presents a study on multi-modal interfaces, issues and problems for establishing communication and interaction between blind and deaf persons. A system-prototype Tyflos-Koufos is proposed in an effort for offering solutions to these challenges.

Multiagent learning attracts much attention in the past few years as it poses very challenging problems. Reinforcement Learning is an appealing solution to the problems that arise to Multi Agent Systems (MASs). This is due to the fact that Reinforcement Learning is a robust and well suited technique for learning in MASs. This paper proposes a multi-agent Reinforcement Learning approach, that uses coordinated actions, which we call strategies and a voting process that combines the decisions of the agents, in order to follow a strategy. We performed experiments to the predator-prey domain, comparing our approach with other multi-agent Reinforcement Learning techniques, getting promising results.

Cerebral aneurysms are weak or thin spots on blood vessels in the brain that balloon out. While the majority of aneurysms do not burst, those that do would lead to serious complications including hemorrhagic stroke, permanent nerve damage, or death. Yet, surgical options for treating cerebral aneurysms carry high risk to the patient. It is vital for the doctors to accurately diagnose aneurysms that have high probabilities of rupturing. In this application, the patient dataset has many attributes, ranging from patient profile to results from diagnostic test and features extracted from brain images. Many of the attributes are discrete and have missing values. The dataset is also highly biased, with 15% unrupture cases and 85% rupture cases. Building a classifier that unerringly predicts the unrupture (rare) class is a challenge. In this paper, we describe a systematic approach to build such a classifier through suitable combination of data mining algorithms. Our approach automatically determines the optimal combination of these algorithms for a dataset. The system has an accuracy of 92% and is currently being deployed at the Huashan Hospital.

We propose in this paper a generic algorithm based on ant colony optimization to solve multi-objective optimization problems. The proposed algorithm is parameterized by the number of ant colonies and the number of pheromone trails. We compare different variants of this algorithm on the multi-objective knapsack problem. We compare also the obtained results with other evolutionary algorithms from the literature.

Sports coaches today have access to a wide variety of information sources that describe the performance of their players. However, despite this great wealth of information, most techniques used to analyse performance require a significant amount of manual processing and continue to rely heavily on input from human experts. In this paper we propose an automated approach to analyse player performance. Specifically, we propose a team benchmarking and concept drift tracking based system that (1) generates adaptive baseline player performance norms, (2) interprets player performance over different time lines and (3) identifies and describes key turning points in player performance. The concept drift technique that we describe uses a combination of overlapping data windows and decision tree based learning to process the data.

In this paper we developed an Inductive Logic Programming (ILP) based framework ExOpaque that is able to extract a set of Horn clauses from an arbitrary opaque machine learning model, to describe the behavior of the opaque model with high fidelity while maintaining the simplicity of the Horn clauses for human interpretations.

In this paper we consider the consistency problem of temporal or spatial qualitive constraint networks. A new encoding making it possible to represent and solve this problem in the framework of the prepositional logic is proposed. The definition of this encoding presupposes the existence of a particular order on the basic relations of the qualitative calculus such as that of the conceptual lattice of the interval algebra of Allen.

A novel character recognition method, called character recognition based on neural network and Dempster-Shafer theory (CRNNDS), is proposed in this paper. The CRNNDS integrates a recurrent neural network (RNN) and Dempster-Shafer (D-S) theory to recognize handwritten characters. It employs an RNN to effectively extract oriented features of a handwritten character and then these features are applied to Dempster-Shafer theory which can powerfully estimate the similarity ratings between a recognized character and sampling characters in the character database. Experimental results demonstrate that the CRNNDS system achieves a satisfying recognition performance.

A soft version of the well-known global constraint AllDifferent has been recently introduced for the Max-CSP framework ([9, 15, 16]). In this paper, we propose to soften AllDifferent in the Weighted CSP framework that is more general. We extend the two semantics of violation proposed in [9]: the first one is based on variables and the second one on the decomposition into a set of binary constraints of difference. For the first semantic, we propose a polynomial algorithm which maintains hyper- arc consistency. For the second one, we prove that checking hyper-arc consistency is an NP-Hard problem. So, we propose to maintain a local consistency using a filtering based on lower bounds computation. Finally, we present some experimental results and draw a few perspectives.

A novel affine invariant region detector based on the 4th differential invariant (DI4) is proposed in this paper. The detector combines scale-space theory with an autocorrelation matrix. Since it is proved that DI4 is a scale- space selection function, feature points and their characteristic scales are first detected by the local maxima of the normalized DI4 over scale-space. Then, the auto-correlation matrices, which are used to describe the affine shapes, are estimated on the characteristic scales of the feature points. The ellipse regions given by the auto-correlation matrices are affine invariant. In order to verify the affine invariance, we build up a simulation experiment to test affine invariance using two single- parameter transforms. The experimental results show the detected regions are invariant to rotation, scale and affine transforms as well as robust to illumination changes.

The precision and reliability of the estimation of the effort of software projects is very important for the competitiveness of software companies. Good estimates play a very important role in the management of software projects. Most methods proposed for effort estimation, including methods based on machine learning, provide only an estimate of the effort for a novel project. In this paper we introduce a method based on machine learning which gives the estimation of the effort together with a confidence interval for it. In our method, we propose to employ robust confidence intervals, which do not depend on the form of probability distribution of the errors in the training set. We report on a number of experiments using two datasets aimed to compare machine learning techniques for software effort estimation and to show that robust confidence intervals for the effort estimation can be successfully built.

In this paper, we propose a new value ordering heuristic for weighted constraint satisfaction problems (WCSP) based on the quality of solutions. The H-quality of an assignment estimates its capacity at occurring in solutions of "good" quality. Experiments using limited discrepancy search on random WCSP instances and CELAR benchmarks show that our value ordering always outperforms MinAC in a significant way.

A considerable amount of attention has been paid to the coalition formation problem to deal efficiently with tasks needing more than one agent. However, little attention has been paid to the problem of monitoring a coalition during the execution by modifying it according to the progress of the accomplishment of the task. In this paper, we consider a coalition of resource-bounded autonomous agents with anytime behavior solving a common complex task. There is no central control component. Agents can observe the effect of the other agents' actions. They can decide whether they should continue to contribute in solving the common task or to stop their contribution and to leave the coalition. This decision is made in a distributed way. The objective is to avoid the waste of resources and time by using the same coalition along the task accomplishment while some agents become not necessary to pursue the accomplishment of the task. We formalize this decentralized decision making problem as a DEC-MDP.

Constraint programming offers a declarative approach to solving problems modeled as constraint satisfaction problems (CSPs). However, the precise specification of a set of constraints is sometimes not available, but may have to be learned, for instance, from a set of examples of its solutions and non-solutions. In general, one may wish to learn generalized CSPs involving classical, fuzzy, weighted or probabilistic constraints, for example. This paper introduces a unifying framework for CSP learning. The framework is generic in that it can be instantiated to obtain specific formulations for learning classical, fuzzy, weighted or probabilistic CSPs. In particular, a new formulation for classical CSP learning, which minimizes the number of examples violated by candidate CSPs, is obtained by instantiating the framework. This formulation is equivalent to a simple pseudo-boolean optimization problem, thus being efficiently solvable using many optimization tools.

Modern software development is currently seeking new paths to improve quality and meet time and cost constraints. Reuse of existing software components is considered one of these paths. However, this process experiences significant problems related to efficiently maintaining component repositories, and, moreover, providing the means to discover and retrieve the most suitable ones. This paper aims to provide a methodology to improve the component-based software development process. Specifically, its objective is to introduce an approach that reduces the time to locate suitable software components. The suggested methodology meets the requirements for the efficient searching of components in repositories and also addresses the need for adequate retrieval of the most suitable software components based on the needs of developers. To achieve this we employ a combination of partitional clustering algorithms borrowed from the field of computational intelligence and fuzzy logic thus creating a subset of the available components that are most suitable to the developers' preferences.

Practical optimization problems often have objective functions that cannot be easily calculated. As a result, comparison-based algorithms that solve such problems use comparison functions that are imperfect (i.e. they may make errors). Machine learning algorithms that search for game-playing programs are typically imperfect comparison algorithms. This paper presents M2ICAL, an algorithm analysis tool that uses Monte Carlo simulations to derive a Markov chain model for imperfect comparison algorithms. Once an algorithm designer has modeled an algorithm using M2ICAL as a Markov chain, it can be analyzed using existing Markov chain theory. Information that can be extracted from the Markov chain include the estimated solution quality after a given number of iterations; the standard deviation of the solutions' quality; and the time to convergence.

In the field of medical image processing, one of the current interests is the automatic segmentation of abdominal CT images, which is the first and fundamental step both before any automatic pathology detection and to compute the 3D abdominal organ volumes. In this paper we propose our fully automatic system that segments liver and its blood vessels, kidneys, spleen, ribs and spine. The overall system has been evaluated on the data of 40 patients, obtaining a good assessment both by visual inspection by three experts, and by comparing the computed results to the manually traced boundaries.

Given a document collection, it is often desirable to find the core subset of documents focusing on a specific topic. We propose a new algorithm for this task. Document clustering aims at partitioning the document-term datasets into different groups by optimizing certain objective functions. However, they are not suitable for finding hotspots that are described by a small set of documents with few tightly coupled terms. In this paper we propose a novel hot spot finding algorithm, DCC (Dense Concept Clustering) in document collections. DCC can extract distinct small topics with most representative documents and words simultaneously. The hotspots are dense bicliques in binary document-word matrices and they can be discovered sequentially one at a time using the generalized Motzkin-Straus formalism. The representative documents and words are tightly correlated for concept descriptions. Experiments on real document datasets show the effectiveness of the proposed algorithm.

In e-learning initiatives, sequencing problem concerns arranging a particular set of learning units in a suitable succession for a particular learner. Sequencing is usually performed by instructors, who create general and ordered series rather than learner personalized sequences. This paper proposes an innovative intelligent technique for learning object automated sequencing using particle swarms. E-learning standards are promoted in order to ensure interoperability. Competencies are used to define relations between learning objects within a sequence, so that the sequencing problem turns into a permutation problem and AI techniques can be used to solve it. Particle Swarm Optimization (PSO) is one of such techniques and it has proven with good performance solving a wide variety of problems. An implementation of the PSO, for learning object sequencing, is presented and its performance in a real scenario is discussed.

In this paper we present a new approach for curve clustering designed for analysis of spatiotemporal data. Such kind of data contains both spatial and temporal patterns that we desire to capture. The proposed methodology is based on regression and Gaussian mixture modeling and the novelty of the herein work is the incorporation of spatial smoothness constraints in the form of a prior for the data labels. This enables the proposed model to take into account the underlying property of spatiotemporal data that spatially adjacent data points most likely should belong to the same cluster. A maximum a posteriori Expectation Maximization (MAP-EM) algorithm is used for learning this model. We present numerical experiments with simulated data where the ground truth is known in order to assess the value of the introduced smoothness constraint, and also with real cardiac perfusion MRI data. The results are very promising and demonstrate the value of the proposed constraint for analysis of such data .

Agent-based e-learning systems allow the interaction between students and e-learning Web sites, providing students with useful suggestions about the available educational resources. In these systems, generally each student is monitored by a student agent, while each e-learning site is associated with a site agent. However, these systems are not able to support students which exploit different devices as PC, palmtop, cellular, etc. multi-agent system handling adaptivity (MASHA) is a recent multi-agent system which appears as a promising candidate to overcome such a limitation. However, in the case of large agent communities, the tasks of both the student and the site agent in MASHA can result significantly heavy. To face this issue, we propose in this paper an extension of the MASHA system, called MASHA-EL, appositely conceived for supporting e-learning. In this system, a student that exploits a given device is provided with a device agent, and each e-learning Web site is associated, in its turn, with a teacher agent. When a student visits the e-Learning site of the teacher agent using a given device, the teacher agent provide him with useful recommendations also considering the exploited device. We present some experimental results that compares the performances of MASHA-EL with other well- known agent-based recommenders, and that show significant advantages obtained by MASHA-EL in terms of recommendation effectiveness and reduction of time costs.

Software reuse is essential for improving the productivity and quality of software projects. One of the key issues to promote the adoption of software reuse in companies is the development of effective repositories of software components. It is also very important to have good methods for searching and retrieval of the components. Clustering techniques can help by providing a visualization of the repository of software components as well as in helping to refine the searches by grouping together similar components. In this paper we quantitatively compare two clustering techniques, namely, self-organizing maps (SOM) and growing hierarchical SOM (GHSOM) for clustering a repository of classes from a Java API for building mobile systems. The performance measure was the quantization error. The simulations have shown that GHSOM outperforms SOM in these tasks. GHSOM is more suitable for this task because it is a constructive technique, which is an advantage in tackling the growth of the repository of software components.

The Software Project Management Antipattern Intelligent System (PROMAISE) is proposed as a Web-enabled knowledge-base framework that uses antipattern OWL ontologies in order to provide intelligent and up to date advice to software project managers regarding the selection of appropriate antipatterns in a software project. Antipatterns provide information on commonly occurring solutions to problems that generate negative consequences. These mechanisms are documented using informal paper based structures that do not readily support knowledge sharing and reuse. Antipattern OWL ontologies can be used to build a dynamic antipattern knowledge base, which can update itself automatically. This will allow the accessibility and transferability of up-to-date computer-mediated software project management knowledge to software project managers by encoding antipatterns into computer understandable ontologies. PROMAISE can function with this knowledge base in order to assist software project managers in the process of selecting applicable antipatterns.

Modern software development is currently seeking new paths to improve quality and meet time and cost constraints. Reuse of existing software components is considered one of these paths. However, this process experiences significant problems related to efficiently maintaining component repositories, and, moreover, providing the means to discover and retrieve the most suitable ones. This paper aims to provide a methodology to improve the component-based software development process. Specifically, its objective is to introduce an approach that reduces the time to locate suitable software components. The suggested methodology meets the requirements for the efficient searching of components in repositories and also addresses the need for adequate retrieval of the most suitable software components based on the needs of developers. To achieve this we employ a combination of partitional clustering algorithms borrowed from the field of computational intelligence and fuzzy logic thus creating a subset of the available components that are most suitable to the developers' preferences. © 2007 IEEE.

Cooperative coevolution is becoming increasingly popular in solving difficult optimization problems. Its performance to solve the problems is influenced by many algorithm decisions. In this paper, a self-adaptive mutation operator "guided mutation" is proposed. The basic idea behind guided mutation is to maintain searching directions and searching step sizes at individual level, and these two strategy parameters are adaptively updated. Guided mutation is adopted in cooperative coevolutionary algorithm and its performance on the common test problems is compared. Experimental results show that guided mutation can improve cooperative coevolution in solving some problem domains. The reasons behind the differences in the performance of the various cooperative coevolutions are also discussed.

With the completion of the human genome and the entrance into the post-genomic era, translational research rises as a major need. In this paper, we present a knowledge discovery workflow (KDw) and its utilization in the context of clinico-genomic trials. KDw aims towards the discovery of 'evidential' correlations between patients' genomic and clinical profiles. Application of KDw on a real-world clinico-genomic (breast cancer) study demonstrates the reliability, efficacy, and efficiency of the approach.

The ability of an algorithm to accurately estimate the parameters of the geometric transformation which aligns two image profiles even in the presence of photometric distortions can be considered as a basic requirement in many computer vision applications. Projective transformations constitute a general class which includes as special cases the affine, as well as the metric subclasses of transformations. In this paper the applicability of a recently proposed iterative algorithm, which uses the Enhanced Correlation Coefficient as a performance criterion, in the projective image registration problem is investigated. The main theoretical results concerning the iterative algorithm and an efficient approximation that leads to an optimal closed form solution (per iteration) are presented. Furthermore, the performance of the iterative algorithm in the presence of nonlinear photometric distortions is compared against the leading Lucas-Kanade algorithm by performing numerous simulations. In all cases the proposed algorithm outperforms the Lucas-Kanade algorithm in convergence speed and robustness against photometric distortions under ideal and noisy conditions.

Mobile robot localization is taken into account as one of the most important topics in robotics. In this paper, the localization problem is extended to the cases in which estimating the position of multi robots is considered. To do so, the joint probabilistic data association filter (JPDAF) approach is applied for tracking the position of multiple robots. To characterize the motion of each robot, two models are used. First, a simple near constant velocity model is considered and then a variable velocity model is applied for tracking. This improves the performance when the robots change their velocity and conduct maneuvering movements. This issue gives an advantage to explore the movement of the manoeuvring objects which is common in many robotics problems such as soccer or rescue robots. Simulation results show the efficiency of the JPDAF algorithm in tracking multiple mobile robots with maneuvering movements.

The multiplication of very high resolution (spatial or spectral) remote sensing images appears to be an opportunity to identify objects in urban and periurban areas. The classification methods applied in the object-oriented image analysis approach could be based on the use of domain knowledge. A major issue in these approaches is domain knowledge formalization and exploitation. In this paper, we propose a recognition method based on an ontology which has been developed by experts of the domain. In order to give objects a semantic meaning, we have developed a matching process between an object and the concepts of the ontology. Experiments are made on a Quickbird image. The quality of the results shows the effectiveness of the proposed method.

Online buyer coalition formation problem is an application of e-commerce and distributed agent technology. Most existing works in this topic involve social utility based approaches that assume the agents' utilities to be transferable. However, we argue that there are situations where the transferable utility model is not well suited for the problem, and that coalition stability is a more important solution concept than social utility. In this paper, we study the problem from a non-transferable utility approach where the focus is on achieving stable solutions in term of the core and Pare to efficiency. A new distributed mechanism is proposed where agents are allowed to propose incremental improvements toward a stable solution. We show by experiment that our mechanism is able to reach core-stable solutions in over 99% of the cases, which suggests a big improvement over the existing approaches.

This paper presents the first implementation of COMMA, a new solver independent language for modeling constraint- based problems. The combination of a constraint language with an object-oriented framework represents the base of the core of COMMA. Extension capabilities have also been included with the aim of tackling a wide range of applications from combinatorial to continuous problems. A COMMA compiler has been implemented through a three layered architecture including a dynamic parsing system for handling efficiently the mapping process. In particular, COMMA models can be translated to different solvers, currently to ECLiPSe and Gecode/J.

The amount and heterogeneity of data in biomedical research, notably in interdisciplinary fields, requires new methods for the collection, presentation and analysis of information. Important data from laboratory experiments as well as patient trials are available but come out of distributed resources. The Charite - University Hospital Berlin has established together with the German Research Foundation (DFG) a new information service centre for kidney diseases and transplantation (Open European Nephrology Science Centre - OpEN.SC). Beside a collaborative aspect to create new research groups every single partner or institution of this science information centre making his own data available is allowed to search the whole data pool of the various involved centres. For an efficient retrieval we develop an intelligent catalogue using principles of knowledge management techniques coming from case based reasoning. The complete system is based on a service-oriented architecture (SOA) with main and auxiliary modules arranged in four layers. Here we introduce a first case study with data coming from the Web-based electronic patient record TBasecopy. The paper is organized as follows: After giving an introduction about the situation of available medical data in nephrology and pathology we introduce the environment of the whole system and come then to our first experiments with a case-based reasoning method.

This paper presents a semi-automatic approach for confounding-aware subgroup discovery: We present a method that provides the means for detecting potentially confounded subgroup patterns, other unconfounded relations, and/or patterns that are affected by effect- modification. Since there is no purely automatic test for confounding, the discovered relations are subsequently presented to the user in a semi-automatic approach. Furthermore, we show how to utilize (causal) domain knowledge for improving the results of the algorithm, since confounding is itself a causal concept. The applicability and benefit of the presented technique is illustrated by examples from a case-study in the medical domain.

Clustering is particularly useful in problems where there is little prior information about the data under analysis. This is usually the case when attempting to evaluate a software system's maintainability, as many dimensions must be taken into account in order to reach a conclusion. On the other hand partitional clustering algorithms suffer from being sensitive to noise and to the initial partitioning. In this paper we propose a novel partitional clustering algorithm, k-Attractors. It employs the maximal frequent itemset discovery and partitioning in order to define the number of desired clusters and the initial cluster attractors. Then it utilizes a similarity measure which is adapted to the way initial attractors are determined. We apply the k-Attractors algorithm to two custom industrial systems and we compare it with WEKA 's implementation of K-Means. We present preliminary results that show our approach is better in terms of clustering accuracy and speed.

Lookahead is a well-known constraint propagation technique for DPLL-based SAT and answer set solvers. Despite its space pruning power, it can also slow down the search, due to its high overhead. In this paper, this twofold effect is analyzed. On one side, we give characterizations of the problems for which the cause for the reduction of search efficiency shows clearly. On the other we show that problem instances that lie in the phase transition regions often significantly benefit from the use of lookahead. Our analysis leads to a proposal of adaptive lookahead, which performs lookahead according to the learned information during the search. Adaptive lookahead is implemented in one of the best-known answer set solvers, smodels. Our experiments show that adaptive lookahead adapts well to different search environments it is going through.

This paper proposes a novel distributional-based approach towards the face retrieval problem. Face features are initially extracted in the frequency domain via Gabor filtering, producing a large number of Gabor jets. The Neural-Gas vector quantizer is used to extract representative samples of the multivariable face distribution. In this way, only a small amount of Gabor jet signatures is utilized. Each face image is then represented as a distribution of a few signatures in the frequency space, containing all the important information. The similarity between two images is finally assessed by comparing the corresponding distributions directly in the frequency space using the multivariate Waid-Wolfowitz test (WW-test), a non- parametric statistical test dealing with the multivariate "Two-Sample Problem". Experimental results drawn from a standard collection of face-images show a significantly improved performance relative to other typical methods.

This paper presents a methodology for the real-time extraction of readers' emotional state from documents as well as the representation of emotionally annotated documents into acoustic modality. Using the Self Assessment Manikin Test we performed preliminary experiments on the extraction of Pleasure - Arousal - Dominance (P.A.D.) annotation rules that the documents evoke to the readers. The rules are used in an automated procedure that assigns text's formatting and structure of documents (meta-data) to emotional state values. During the vocalization of documents, these values, and consequently the documents' meta-data, are carried by different expressions of speech using text-to-speech synthesis. The proposed system architecture is language independent and content-free.

This paper proposes a dynamic software testing framework, which is able to analyse the source code of a program, create the necessary data structures for automatic testing, such as control flow graphs, and generate a near to optimum set of test cases with reference to a test coverage criterion. The framework consists of two sub-systems: the first is a program analysis system that identifies the type of statements and the complexity of conditions, performs analysis of variables, extracts code paths and creates the control flow graph (CFG) of the program under testing. The second is a test system that uses the CFG for automatically generating test data based on evolutionary computing. The latter system utilises a specially designed genetic algorithm to produce the set of test cases satisfying the selected coverage criterion. The efficacy and performance of the proposed testing approach is assessed and validated using a variety of sample programs.

Recently, some 2D based algorithms were proposed for image classification. In this paper, we propose two-dimensional logistic regression for two-class problem. This method has an advantage of memory saving and better performance than other 2D based approaches. The experimental results show two dimensional logistic regression is superior to one dimensional logistic regression and other popular two-dimensional methods.

This paper describes the process of learning taxonomic relations automatically from Modern Greek economic corpora. Supervised learning (Decision trees, Support Vector Machines, Meta-learning) is applied to economic term pairs; each pair is represented through a set of statistical, semantic and syntactic features. The resulting set of feature-value vectors presents a high imbalance in the class distribution, due to the large number of term pairs that do not present a direct semantic relation. This problem is addressed using One-sided Sampling, which reduces the number of the majority class instances by removing examples that are noisy, misleading or redundant. The approach makes use of no external resources (merely an economic corpus that is annotated with elementary morphological and phrase chunking information) and limited language-dependent elements to facilitate its portability to other languages and domains. An overall f-measure of 71 % is achieved.

In this paper, a tool that uses an argumentation based decision making framework is proposed for the construction of mutual fund portfolios. The argumentation framework is employed in order to develop mutual funds performance models and to select a small set of mutual funds, which will compose the final portfolio. The knowledge engineering approach and tool development know-how, presented here-in, can be used for the development of other real-world applications using argumentation.