Intelligent Data Engineering and Automated Learning - IDEAL 2007: 8th International Conference, Birmingham, UK, December 16-19, 2007, Proceedings (Lecture Notes in Computer Science, 4881)

This book constitutes the refereed proceedings of the 8th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2007, held in Birmingham, UK, in December 2007. The papers include topical sections on learning and information processing, data mining and information management, bioinformatics and neuroinformatics, agents and distributed systems, financial engineering and modeling, and agent-based approach to service sciences. Title Page Preface Organization Table of Contents Support Function Machines Introduction Support Function Machines Model Learning Algorithm Application Examples Harm Forecast of Horsetail Pine Worms Stock Price Predictions Conclusions Different Bayesian Network Models in the Classification of Remote Sensing Images Introduction Bayesian Networks Bayesian Network Classifiers Remote Sensing Image Classification An Example of Multispectral Data Set Analysis An Example of Hyperspectral Data Set Analysis Conclusions Group Decision Making with Triangular Fuzzy Linguistic Variables Introduction Triangular Fuzzy Linguistic Variables Some Aggregation Operators A Method for Group Decision Making with Triangular Fuzzy Linguistic Variables Illustrative Example Concluding Remarks References Sparse Kernel Modelling: A Unified Approach Introduction A Unified Framework for Data Modelling Orthogonal-Least-Squares Algorithm Sparse Kernel Regression Model Construction Sparse Kernel Classifier Construction Sparse Kernel Density Estimator Construction Empirical Data Modelling Results Conclusions Advanced Forecasting and Classification Technique for Condition Monitoring of Rotating Machinery Introduction Preparation of Inputs for the Predictor Forecasting Technique and Predictor Architecture Classification Technique and Overall Scheme Conclusions and Outlook References Out of Bootstrap Estimation of Generalization Error Curves in Bagging Ensembles Introduction Monte Carlo Ensemble Learning Error Curves for Bagging Ensembles Bias Analysis Out of Bootstrap Estimation Experiments Conclusions An Edit Distance Approach to Shallow Semantic Labeling Introduction Two-Phase Feature-Enhanced String Matching Algorithm Edit Operations Shallow Semantic Labeling as Two-Phase Feature-Enhanced String Matching Experimental Results Conclusion References A Comparison of One-Class Classifiers for Novelty Detection in Forensic Case Data Introduction Related Work Novelty Detection Density Estimation Methods Boundary Methods Reconstruction Methods Combination of Classifiers ROC Analysis Datasets and Methodology The Data Experimental Setup Results and Discussion Conclusion Variational GTM Introduction Generative Topographic Mapping The Original GTM Gaussian Process Formulation of GTM Bayesian GTM Variational GTM Motivation of the Use of Variational Inference A Bayesian Approach of GTM Based on Variational Inference Experiments Experimental Design Robustness of the Variational GTM in the Presence of Noise Data Visualization Using Variational GTM Conclusions Skill Combination for Reinforcement Learning Introduction Skills in Reinforcement Learning Skills Combination Experiments Cat and Mouse Domain Introduction Skill Training Task 1: Get Cheese Task 2: Cheese Storage Compare Task 1 and Task 2 Discussion Conclusion and Future Work References A New Recurring Multistage Evolutionary Algorithm for Solving Problems Efficiently Introduction Recurring Multistage Evolutionary Algorithm Exploration Stage Exploitation Stage Recurring Approach Experimental Studies Conclusions The Effect of Missing Wind Speed Data on Wind Power Estimation Introduction ANFIS Weibull Distributions Results Discussions and Conclusions Exploration of a Text Collection and Identification of Topics by Clustering Introduction Construction of the Vector Space Model Exploratory Analysis of Existing Categories Average Cosine Measures Between Documents MDS Layouts of the Original Categories Identification of Documents Subsets Proposed Approach for Topic Identification Identified Topics for the Original Categories Clustering of the Abstracts and Evaluation Clustering Experiments Identification of Topics for the Clusters Conclusions Asynchronous BCI Control of a Robot Simulator with Supervised Online Training Introduction Methods BCI Experiment Setup Offline Training Online Asynchronous Event Detection A Robot Simulator and Its Specifically Designed Environment Online Training Results Performance of the Asynchronous BCI without Online Training Performance of the Asynchronous BCI with Online Training Conclusion References Fuzzy Ridge Regression with Non Symmetric Membership Functions and Quadratic Models Introduction Fuzzy Linear Regression Use of Quadratic Programming Fuzzy Ridge Regression Examples Conclusions A Subjective and Objective Integrated Method for MAGDM Problems with Multiple Types of Exact Preference Formats Introduction Problem Presentation Constrained Optimization Models Based on Data Matrix and Each Different Exact Preference Format Constrained Optimization Models Integrating Data matrix and All Three Different Preference Structures Concluding Remarks References Energy Saving by Means of Fuzzy Systems Introduction The Problem Definition The Fuzzy Model for Energy Saving and Comfort The Solution Analysis The Multi Agent System The Estimation of the Power Requirement in a Room The Energy Distribution Algorithm Experiments and Results Conclusions A Comparative Study of Local Classifiers Based on Clustering Techniques and One-Layer Neural Networks Introduction The Local Model The Clustering Algorithms Implemented The One-Layer Neural Networks The Construction of the Local Model Results Artificial Data Sets Real Data Sets Comparative Results with Other Methods Conclusions Filter Methods for Feature Selection – A Comparative Study Introduction The Filter Methods Used Relief Correlation-Based Feature Selection, CFS Fast Correlated Based Filter, FCBF INTERACT Synthetic Datasets Experimental Results Conclusions FPGA-Based Architecture for Computing Testors Introduction Algorithms for Computing Testors Proposed Architecture FPGA Implementation and Results Discussion Conclusions References Minimal BSDT Abstract Selectional Machines and Their Selectional and Computational Performance Introduction BSDT Coding/Decoding and Performance Minimal BSDT ASMs and Their Functions Minimal One-Dimensional BSDT Passive ASM Minimal One-Dimensional BSDT Active ASM Optimal Selectional Performance of Minimal BSDT ASMs Discussion and Conclusion References Active Learning for Regression Based on Query by Committee Introduction Active Learning of Real-Valued Functions Numerical Examples Discussion Conclusion Influence of Wavelet Frequency and Orientation in an SVM-Based Parallel Gabor PCA Face Verification System Introduction Face Databases XM2VTS Database FRAV2D Database Our Algorithm Design of Experiments Results and Discussion Performance of Parallel Gabor PCA Versus Other Methods Influence of Gabor Wavelet Frequency and Orientation in a One-Gabor PCA Conclusions References Wrapping the Naive Bayes Classifier to Relax the Effect of Dependences Introduction and Motivation Wrappers for Attribute Subset Selection The Dependency Guided Wrapper Experiments Discussion of Results Conclusions and Future Work Preference Learning from Interval Pairwise Data. A Distance-Based Approach Introduction Group Preference Learning Generating the Group Preference Weights Conclusions Psychometric Functions Within the Framework of Binary Signal Detection Theory: Coding the Face Identity Introduction Disadvantages of SDT Psychometric Functions Advantages of BSDT Psychometric Functions BSDT Face Neural Subspace, FNS BSDT Explanation of Face Identity Experiments Conclusion References Load Forecasting with Support Vector Machines and Semi-parametric Method Introduction Semi-parametric Method Estimation of the Harmonic Components Non-linear Part Numerical Results Conclusion Reproducing Kernel Hilbert Space Methods to Reduce Pulse Compression Sidelobes Introduction Problem Formulation Matched Filter Least Squares and $L_{2P}$ Sidelobe Minimisation RKHS-Based Filter Numerical Results Conclusion Support Kernel Machine-Based Active Learning to Find Labels and a Proper Kernel Simultaneously Introduction Support Kernel Machines Two-Phased SKM-Based Active Learning: LASKM Sampling Strategies for SKM-Based Active Learning Experiments Comparison Between Single Kernel SVMs and SKM Comparison of Sampling Strategy in SKM-Based Active Learning Conclusions and Future Works Making Class Bias Useful: A Strategy of Learning from Imbalanced Data Introduction Related Work Proposed Solutions Data Cleaning Classification and Ranking Experimental Evaluation Data Set AUC Score Comparison with Existing Methods Influence of Class Distribution Summary Detecting Phishing E-mails by Heterogeneous Classification Introduction Phishing Filtering Classification System Naïve Bayes Classifier for the Textual Content of E-mails Rule-Based Classifier for the Non Grammatical Content of E-mails A Classifier Based on an Emulator for the Content of Websites Addressed by the Links Contained in E-mails Empirical Evaluation Evolution of the System Conclusions and Future Work References Load Balancing in Fault Tolerant Video Server Introduction Previous Works System Architecture Algorithm Simulation and Performance Analysis Conclusions Position-Aware String Kernels with Weighted Shifts and a General Framework to Apply String Kernels to Other Structured Data Introduction A Survey of String Kernels Position-Unaware String Kernels Position-Aware String Kernels with Precise Position Matching Position-Aware String Kernels with Weighted Shifts Our Contributions A Framework for Transforming Character-Base String Kernels into Kernels over Other Structured Data Proof of Theorem 1 Proof of Corollary 1 A New Regression Based Software Cost Estimation Model Using Power Values Introduction Related Work Evaluation Criteria Data Set Description Proposed Model Comparison and Analysis of Results Discussion and Conclusion References Visualising and Clustering Video Data Introduction SOM Topographic Products of Experts Comparison with the GTM Visualising and Clustering Voice Data The Data and Pre-processing Experiments Conclusion Neural Network-Based Receiver for Uplink Multiuser Code Division Multiple Access Communication System Introduction Communication System Model Neural Network-Based Receiver Performance Evaluation Conclusions References Evolving Tree Algorithm Modifications Introduction The Evolving Tree Algorithm Considerations About the E-Tree Algorithm Evolving Tree Algorithm Modifications Image Segmentation Application Conclusions A Linear Learning Method for Multilayer Perceptrons Using Least-Squares Motivation The Proposed Algorithm One-Layer Linear Learning: Determining the Weights One-Layer Linear Learning: Determining the Inputs Combining Theorem 1 and Theorem 2 for Linear Learning of a MLP Experimental Results Breast Cancer Wisconsin Wine Discussion Conclusions and Future Work A Discriminative Model Corresponding to Hierarchical HMMs Introduction Related Work HHMMs Representing an HHMM as a DBN HHCRFs Model Parameter Estimation Marginalized Viterbi Algorithm for HHMMs Marginalized Viterbi Algorithm for HHCRFs Experiment Conclusion Finding Unsatisfiable Subformulas with Stochastic Method Introduction Preliminaries Local Search for Finding Unsatisfiable Subformulas Heuristics and Pruning Technique Experimental Results Conclusion A New Efficient Approach in Clustering Ensembles Introduction Clustering Ensembles The Proposed Method Intelligent K-Means Labeling Algorithm Experiments Heuristic Functions in Labeling Complexities Conclusion References An Evolutionary Hyperheuristic to Solve Strip-Packing Problems Introduction Heuristics Based Methods The Evolutionary Hyperheuristic Approach The Genetic Inspired Hyperheuristic: G-SP Tuning Tests Comparison with Low-Level Heuristics Comparison with State-of-the-Art Algorithms Conclusions Statistical Analysis of Sample-Size Effects in ICA Introduction Blind Source Separation Finite-Sample Effects in the Estimation A Test for Kurtosis Pearson's Families Fisher Information Simulations Discussion Conclusion HPGP: An Abstraction-Based Framework for Decision-Theoretic Planning Introduction Planning in Uncertain Environments Graphplan Framework PGraphplan Searching the Probabilistic Planning Graph HIPU - Hierarchical Planner Under Uncertainty Generating Abstractions in HIPU Hierarchical Planning Solver HPGP – Hierarchical Probabilistic Graphplan Empirical Results Conclusions and Future Work References Correction of Medical Handwriting OCR Based on Semantic Similarity Introduction Semantic Model for Correction of OCR Output Task Formulation Models Semantic Similarity for Words Heuristic rules Merging with the Word Classifier Experiments Conclusions Related Work Multiple Classifier Fusion Using $k$-Nearest Localized Templates Introduction Background Conventional Fusion Methods $C$-Means Algorithm Decision Templates $k$-Nearest Localized Templates Estimation of Localized Decision Templates Classification Using $k$-Nearest Localized Templates Experiments Parameter Setting of the $k$-Nearest Localized Templates Classification Results Conclusions References Color Image Segmentation Applied to Medical Domain Introduction The Color Set Back-Projection Algorithm Content-Based Region Query - Experiments and Results Tracking the Time Evolution of the Disease - Experiments and Results Conclusions References Hierarchical Program Representation for Program Element Matching Introduction Beyond Syntax Trees Program Dependence Higraphs Reduced Blocks The Dominance DAG Dependence Higraph Definition Dependence Higraph Construction Dependence Higraph Matching Abilities Conclusions and Future Work A Combination-of-Tools Method for Learning Interpretable Fuzzy Rule-Based Classifiers from Support Vector Machines Introduction The Fuzzy Rule-Based Classifier The Structure of the Fuzzy Rule-Based Classifier Formulation of the Fuzzy Rule-Based Classifier as a Kernel Machine Fuzzy Rule-Based Classifier Based SVM Reduction of the Number of Fuzzy Sets Reduction of the Number of Rules by Orthogonal Transforms Application Example Conclusions An Effective Content-Based Image Retrieval System by Hierachical Segmentation Introduction Hierarchical Multi-resolution Segmentations Scale and Rotation Invariant Features Geometric Invariants Normalized Histogram Similarity Mmetric Image Retrieval Based-Hierarchical Segmentation Experimental Results Conclusion References Knowledge Extraction from Unstructured Surface Meshes Introduction Surface Representation Displacement Measurement Definition Major Properties Knowledge Extraction from Design Data Displacement Analyzis Sensitivity Analysis Modeling and Analyzing Interrelations Summary Clustering with Reinforcement Learning Introduction The Bernoulli Model Simulation Algorithm RL1 Simulation New Algorithm RL2 Algorithm RL3 Simulation A Topology Preserving Mapping RL1 Topology-Preserving Mapping (RL1ToM) RL2 Topology-Preserving Mapping (RL2ToM) Simulation Conclusion Mining Frequent Itemsets in Large Data Warehouses: A Novel Approach Proposed for Sparse Data Sets Introduction Related Work The Proposed Algorithm Mining Frequent Itemsets Experimental Results Experiment 1: Synthetic Data Experiment 2: Real Databases Conclusion References A Sparse Bayesian Position Weighted Bio-Kernel Network Introduction Selecting Support Peptides Position Weighting Biological Interpretation Method Optimisation of Network Weights Optimising Residue Weights Results and Discussion Datasets Results Model Analysis Conclusions Square Penalty Support Vector Regression Introduction Hard Margin SV Regression and Classification Square Penalty SV Regression and Classification Numerical Experiments Conclusions and Discussion Constructing Accurate Fuzzy Rule-Based Classification Systems Using Apriori Principles and Rule-Weighting Introduction Fuzzy Rule-Based Classification Systems Rule-Base Construction Generating Rules with 1 or 2 Antecedents Generating Rules of Higher Dimensions The Proposed Method for Rule Weighting Finding the Optimal Decision Boundaries Experimental Results Artificial Data Real-Life Data Sets Conclusions References Visualization of Topology Representing Networks Introduction Topology Representing Network Based Mapping Algorithms Topology Representing Network Geodesic Nonlinear Projection Neural Gas Topology Representing Network Map Analysis of the Topology Representing Network Based Mapping Methods Mapping Quality Analysis of the Methods Conclusion The Outer Impartation Information Content of Rules and Rule Sets Introduction The Outer Impartation Information Content of Rules The Outer Impartation Information Content of Rule Sets Experiment An Engineering Approach to Data Mining Projects Introduction Data Mining Process Models Software Engineering Process Models IEEE STD 1074 ISO 12207 Unification of IEEE STD 1074 and ISO 12207 SE Process Model vs. CRISP-DM Life Cycle Selection Process Project Management Processes Development-Oriented Processes Integral Processes A Process Model for Data Mining Engineering Conclusions Classifying Polyphonic Melodies by Chord Estimation Based on Hidden Markov Model Motivation Melody and Polyphony Features for Melody Classification Hidden Markov Model HMM for Melody Classification Experimental Results Preliminaries Results Discussion Conclusion Elastic Non-contiguous Sequence Pattern Detection for Data Stream Monitoring Introduction Related Work Non-contiguous Subsequence Problem Formalization ETC-NSP Detection in Streaming Environment Proposed Method Naive Approach Minimal Variance Matching Pruning Methods in Streaming Environment Triangle Inequality Lower Bound Distance Complexity Analysis Experiment Evaluation Data Set Capture of Non-contiguous Sequence Patterns Pruning Power Conclusion Joint Cutoff Probabilistic Estimation Using Simulation: A Mailing Campaign Application Introduction Campaign Design with One Product Using Simulation and Data Mining for a Campaign Design with More Than One Product Experiments with N Products Experimental Settings Experimental Results Conclusion Segmentation and Annotation of Audiovisual Recordings Based on Automated Speech Recognition Introduction Related Work Algorithm Description Evaluation Measurement Comparison of Algorithms Result Conclusion Mining Disjunctive Sequential Patterns from News Stream Motivation Disjunctive Patterns Problem Definition Algorithm Experimental Results Preliminaries Results Discussion Conclusion A New Dissimilarity Measure Between Trees by Decomposition of Unit-Cost Edit Distance Introduction Preliminaries Unit-Cost Edit Distance Tree Inclusion Flaw of Unit-Cost Edit Distance Our Dissimilarity Measures Decomposition of Unit-Cost Edit Distance Definitions of Our Dissimilarity Measures Related Works Application to Noisy Subsequence Tree Recognition Application to Classification of XML Documents Experimental Results Conclusion Optimizing Web Structures Using Web Mining Techniques Introduction Related Work Overview Web Structure Mining Web Log Mining Analysis Evaluation Summary and Conclusions A Collaborative Recommender System Based on Asymmetric User Similarity Introduction Collaborative Recommender Systems A Recommender System Based on Asymmetric User Similarity Asymmetric Users’ Similarity Predicted Scores Deployment Preliminary Evaluation Final Remarks References Stop Wasting Time: On Predicting the Success or Failure of Learning for Industrial Applications Introduction Bias-Variance Decomposition: A Review Basic Definitions of Bias, Variance and Noise Kohavi and Wolpert's Definition of Bias and Variance Bias as an Upper Limit on Accuracy Prediction Methodology Experimental Methodology Results and Discussion Conclusion Parallel Wavelet Transform for Spatio-temporal Outlier Detection in Large Meteorological Data Introduction Parallel Spatio-temporal Outlier Detection Calculation of Daubechies $D4$ Wavelet Transform Parallel Daubechies Fast Wavelet Transform Parallel Outlier Mining Experimental Result Spatial Analysis Parallel Algorithm Conclusions A Tool for Web Usage Mining Introduction Web Log Processing User Session Identification Web Page Classification Transaction Identification Model Generation Experiments Conclusions An Algorithm to Mine General Association Rules from Tabular Data Introduction Background and Related Work Terminology and Notation Finding Supports of General Itemsets The MGR Algorithm Mining Half General Itemsets Mining General Itemsets Time Complexity and Memory Management of the MGR Algorithm Experimental Results Requirements of the Algorithm Experimental Results Conclusions References Intrusion Detection at Packet Level by Unsupervised Architectures Introduction A Connectionist-Based Framework for IDS Unsupervised Methods for IDS Implementation Vector Quantization Self-organizing Maps Auto-associative Back-Propagation Networks Unsupervised Connectionist Methods for IDS’s The Feature Set Set-Up of the Unsupervised Connectionist Models Experimental Results VQ Paradigm Self-organizing Map AABP Paradigm References Quality of Adaptation of Fusion ViSOM Introduction Quality Measures for Topology Preserving Models The ViSOM Learning Algorithm Quality Measures Topology Preserving Mapping Fusion Use of the Ensemble Meta-algorithm Fusion Variants Under Study Performance Experiments Conclusions and Future Work References Classification Based on the Trace of Variables over Time Introduction Representing Data Histories by Interval Sequences Related Work Increasing the Expressiveness Choice of the Classifier Sets of Relations Evaluation Conclusion Extracting Meaningful Contexts from Mobile Life Log Introduction Related Works Proposed Method Log Collection Log Preprocessing and Context Generation Landmarks Derived from Contexts Data Mining by KeyGraph Experimental Results Log Collection Log Preprocessing and Context Generation Inference from Contexts Data Mining Summary and Future Work References Topological Tree Clustering of Social Network Search Results Introduction Visual Representation of Retrieved Social Content Searching for Profiles and Groups in a Social Network The Importance of Clustering and Topology The Topological Tree Method Overview of the Method Growing Chains and Topological Tree Method Results and Discussions Comparison and Discussion Conclusion and Future Work References A Framework to Analyze Biclustering Results on Microarray Experiments Introduction Bicluster Visualization Microarray Data Visualizations Bubble Map TRN Visualization Linked Visualizations Case Study Example Dataset Objectives Overview Bicluster-Oriented Analysis Gene and Condition-Oriented Analysis Conclusion and Future Work Methods to Bicluster Validation and Comparison in Microarray Data Introduction Bicluster Structure Bicluster Classification Coherence Measures Validation Indices Biological External Indices Non-biological External Indices Internal Indices Relative Indices Application Algorithms Data Sets Methods Results Conclusions and Future Work Capturing Heuristics and Intelligent Methods for Improving Micro-array Data Classification Introduction Related Work The Learning Strategy Experiments Discussion and Concluding Remarks References Classification of Microarrays with kNN: Comparison of Dimensionality Reduction Methods Introduction Dimensionality Reduction Principal Component Analysis (PCA) Random Projection (RP) Partial Least Squares (PLS) Information Gain (IG) Empirical Study Data Sets Experimental Setup Experimental Results Concluding Remarks Protein Data Condensation for Effective Quaternary Structure Classification Introduction Related Work Classification Method Experiments Concluding Remarks $PINCoC$: A Co-clustering Based Approach to Analyze Protein-Protein Interaction Networks Introduction Notation and Problem Definition Algorithm Description Experimental Validation Concluding Remarks Discovering α–Patterns from Gene Expression Data Introduction Definitions Algorithm Method and Experimental Results Conclusions Biclusters Evaluation Based on Shifting and Scaling Patterns Introduction Shifting and Scaling Patterns Unconstrained Optimization Techniques Formulation of the Problem Experiments Conclusions A Deterministic Model to Infer Gene Networks from Microarray Data Introduction Related Work Approach Building Regression Trees Building the Gene Network Experiments Conclusions Profiling of High-Throughput Mass Spectrometry Data for Ovarian Cancer Detection Introduction Methods Data Preprocessing SNEO Based Peak Detection Peak Calibration Peak Extraction Random Forest Based Classification Results Conclusion Adapting Machine Learning Technique for Periodicity Detection in Nucleosomal Locations in Sequences Introduction Related Work Periodicity Detection Experimental Evaluation Conclusions References Analysis of Tiling Microarray Data by Learning Vector Quantization and Relevance Learning Introduction The Classification Problem Features for Classification Data Set and Validation Procedure Fixed Metrics Classifiers Class Conditional Means Learning Vector Quantization Adaptive Metrics Classifiers Relevance Learning Vector Quantization Discussion and Outlook Discriminating Microbial Species Using Protein Sequence Properties and Machine Learning Introduction Methodology Results Discussion Automatic Prognostic Determination and Evolution of Cognitive Decline Using Artificial Neural Networks Introduction Data Environment Neural Approximation Sanger Network Extension for Missing Data Treatment Results and Discussion Patients Prognostic Temporary Evolution of Patients Conclusions SCSTallocator: Sized and Call-Site Tracing-Based Shared Memory Allocator for False Sharing Reduction in Page-Based DSM Systems Introduction Related Works $SCSTallocator$: Combining the Sized and $CSTallocator$ Performance Evaluation Experimental Environments Experimental Results Analysis of Space Efficiency Conclusions and Future Works References A Hybrid Social Model for Simulating the Effects of Policies on Residential Power Consumption Introduction Multi-agent Social Simulations Hybrid Social Model Based on Multi-agent Residential Power Demand-Supply Chain Residential Power Consumption Forecast Model Artificial Residents Society Simulation of Residential Power Consumption RECMAS Architecture Agent Role Simulation Procedure Experiment Analysis Calibration and Validation of the Model Simulation Experiment Scenarios Consumer Types Simulation Analysis Results Conclusion On Intelligent Interface Agents for Human Based Computation Introduction Intelligent Interface Agent Systems Interface Agents for Human Based Computation Agents for Human Based Computation Reviewing Human Based Computation Applications Google Image Labeler Spam Detection Conclusions Reverse Engineering an Agent-Based Hidden Markov Model for Complex Social Systems Introduction Overview of Agent-Based Hidden Markov Model State of the Society State Transitions Communications Learning Process Learning from Communications Learning from Group Structure Learning from Group Evolution Experiments Results on Synthetic Data Results on Real Data Conclusions Effects of Neighbourhood Structure on Evolution of Cooperation in N-Player Iterated Prisoner’s Dilemma Introduction Background Methodology The Cellular Automata Model Strategy Representation Strategy Choice Experiments and Results Experimental Results on Triangular Neighbourhood Structure Experimental Results on Rectangular Neighbourhood Structure Experimental Results on Random Pairing Structure Conclusion References Interface Agents’ Design for a DRT Transportation System Using PASSI Introduction Related Work Flexible Public Transport Services PASSI Methodology The Agent-Based Transportation System The Agent Interaction The Client Agent Conclusions A Multi-agent System Approach to Power System Topology Verification Introduction The Theoretical Background of the Approach A Power System Model Utilized Relationships A Classical Approach to Power System Analyses The Utilized Idea of Power System Analyses A Description of the Approach The Node Agents The Branch Agents Conclusions References A System for Efficient Portfolio Management Introduction Portfolio Theory Portfolio Selection Asset Analysis Computing the Historical Payoff Derivation of Efficient Border Computing the Optimum Portfolio Analysis of Risk Evaluation of Optimum Portfolio System Description Specification of the Asset Set Analysis of Assets Through Estimation of Expected Payoffs, Variances and Covariances Determination of the Efficient Border Neural Network Classification Experiments Experimental Results Conclusions Partitioning-Clustering Techniques Applied to the Electricity Price Time Series Introduction Partitioning-Clustering Techniques K-Means Clustering Technique Expectation Maximization Results K-Means Results EM Results Conclusions Time-Series Prediction Using Self-Organising Mixture Autoregressive Network Introduction Methodology Lampinen and Oja's Self-organising AR Models Self-organising Mixture AR Models (SOMAR) Experimental Results Mackey-Glass Data Foreign Exchange Rate Data Conclusions Adjusting the Generalized Pareto Distribution with Evolution Strategies – An application to a Spanish Motor Liability Insurance Database Introduction Estimation of Generalized Pareto Distribution Evolution Strategies Problem Resolution with Classical Techniques Problem Resolution with Evolution Strategies Conclusion and Discussion References Independent Factor Reinforcement Learning for Portfolio Management Introduction Proposed System Factor Construction Module Weight Conversion Module RL Module Portfolio Optimization Overlay Another RL-Based Portfolio Management System Experimental Results and Analysis Experimental Results Analysis on Portfolio Formation Conclusion and Future Work Discrete Time Portfolio Selection with Lévy Processes Introduction Discrete Time Portfolio Selection with Subordinated Lévy Processes The Discrete Time Portfolio Selection Problem A Comparison Among Lévy Dynamic Strategies Concluding Remarks References Analyzing the Influence of Overconfident Investors on Financial Markets Through Agent-Based Model Introduction Model Negotiable Assets in the Market Modeling Investor Behavior Equity Price Forecasti This book constitutes the refereed proceedings of the 8th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2007, held in Birmingham, UK, in December 2007. The 170 revised full papers presented were carefully reviewed and selected from more than 270 submissions. The papers are organized in topical sections on learning and information processing, data mining and information management, bioinformatics and neuroinformatics, agents and distributed systems, financial engineering and modeling, agent-based approach to service sciences, as well as neural-evolutionary fusion algorithms and their applications