Machine Learning: ECML 2004 15th European Conference on Machine Learning, Pisa, Italy, September 20-24, 2004, Proceedings

This book constitutes the refereed proceedings of the 15th European Conference on Machine Learning, ECML 2004, held in Pisa, Italy, in September 2004, jointly with PKDD 2004.
The 45 revised full papers and 6 revised short papers presented together with abstracts of 5 invited talks were carefully reviewed and selected from 280 papers submitted to ECML and 107 papers submitted to both, ECML and PKDD. The papers present a wealth of new results in the area and address all current issues in machine learning.

Invited Papers.- Random Matrices in Data Analysis.- Data Privacy.- Breaking Through the Syntax Barrier: Searching with Entities and Relations.- Real-World Learning with Markov Logic Networks.- Strength in Diversity: The Advance of Data Analysis.- Contributed Papers.- Filtered Reinforcement Learning.- Applying Support Vector Machines to Imbalanced Datasets.- Sensitivity Analysis of the Result in Binary Decision Trees.- A Boosting Approach to Multiple Instance Learning.- An Experimental Study of Different Approaches to Reinforcement Learning in Common Interest Stochastic Games.- Learning from Message Pairs for Automatic Email Answering.- Concept Formation in Expressive Description Logics.- Multi-level Boundary Classification for Information Extraction.- An Analysis of Stopping and Filtering Criteria for Rule Learning.- Adaptive Online Time Allocation to Search Algorithms.- Model Approximation for HEXQ Hierarchical Reinforcement Learning.- Iterative Ensemble Classification for RelationalData: A Case Study of Semantic Web Services.- Analyzing Multi-agent Reinforcement Learning Using Evolutionary Dynamics.- Experiments in Value Function Approximation with Sparse Support Vector Regression.- Constructive Induction for Classifying Time Series.- Fisher Kernels for Logical Sequences.- The Enron Corpus: A New Dataset for Email Classification Research.- Margin Maximizing Discriminant Analysis.- Multi-objective Classification with Info-Fuzzy Networks.- Improving Progressive Sampling via Meta-learning on Learning Curves.- Methods for Rule Conflict Resolution.- An Efficient Method to Estimate Labelled Sample Size for Transductive LDA(QDA/MDA) Based on Bayes Risk.- Analyzing Sensory Data Using Non-linear Preference Learning with Feature Subset Selection.- Dynamic Asset Allocation Exploiting Predictors in Reinforcement Learning Framework.- Justification-Based Selection of Training Examples for Case Base Reduction.- Using Feature Conjunctions Across Examples for Learning Pairwise Classifiers.- Feature Selection Filters Based on the Permutation Test.- Sparse Distributed Memories for On-Line Value-Based Reinforcement Learning.- Improving Random Forests.- The Principal Components Analysis of a Graph, and Its Relationships to Spectral Clustering.- Using String Kernels to Identify Famous Performers from Their Playing Style.- Associative Clustering.- Learning to Fly Simple and Robust.- Bayesian Network Methods for Traffic Flow Forecasting with Incomplete Data.- Matching Model Versus Single Model: A Study of the Requirement to Match Class Distribution Using Decision Trees.- Inducing Polynomial Equations for Regression.- Efficient Hyperkernel Learning Using Second-Order Cone Programming.- Effective Voting of Heterogeneous Classifiers.- Convergence and Divergence in Standard and Averaging Reinforcement Learning.- Document Representation for One-Class SVM.- Naive Bayesian Classifiers for Ranking.- Conditional Independence Trees.- Exploiting Unlabeled Data in Content-BasedImage Retrieval.- Population Diversity in Permutation-Based Genetic Algorithm.- Simultaneous Concept Learning of Fuzzy Rules.- Posters.- SWITCH: A Novel Approach to Ensemble Learning for Heterogeneous Data.- Estimating Attributed Central Orders.- Batch Reinforcement Learning with State Importance.- Explicit Local Models: Towards “Optimal” Optimization Algorithms.- An Intelligent Model for the Signorini Contact Problem in Belt Grinding Processes.- Cluster-Grouping: From Subgroup Discovery to Clustering.