Advances in Intelligent Data Analysis VII

Weareproudtopresenttheproceedingsoftheseventhbiennialconferenceinthe Intelligent Data Analysis series. The conference took place in Ljubljana, Slo- nia, September 6-8, 2007. IDA continues to expand its scope, quality and size. It started as a small side-symposium as part of a larger conference in 1995 in Baden-Baden(Germany).It quickly attractedmoreinterest in both submissions and attendance as it moved to London (1997) and then Amsterdam (1999). The next three meetings were held in Lisbon (2001), Berlin (2003) and then Madrid in 2005. The improving quality of the submissions has enabled the organizers to assemble programs of ever-increasing consistency and quality. This year we madea rigorousselectionof33papersoutofalmost100submissions.Theresu- ing oral presentations were then scheduled in a single-track, two-and-a-half-day conference program, summarized in the book that you have before you. In accordance with the stated IDA goal of bringing together researchers from diverse disciplines, we believe we have achieved an excellent balance of presentationsfromthemoretheoreticalbothstatisticalandmachinelearning to the more application-oriented areas that illustrate how these techniques can beusedinpractice.Forexample,theproceedingsincludepaperswiththeoretical contributions dealing with statistical approaches to sequence alignment as well as papers addressing practical problems in the areas of text classi?cation and medical data analysis. It is reassuring to see that IDA continues to bring such diverse areas together, thus helping to cross-fertilize these ?elds.

Autorentext
Prof. Dr. Nada Lavrac heads the Department of Knowledge Technologies at the Jo ef Stefan Institute in Ljubljana. She is the author and editor of several books and proceedings in the field of data mining and machine learning, and she has chaired or served on the boards of the main related journals and conferences. Her research interests include machine learning, data mining, and inductive logic programming, and related applications in medicine, public health, bioinformatics, and the management of virtual enterprises. In 1997 she was awarded the Ambassador of Science of Slovenia prize, and in 2007 she was elected as an ECCAI Fellow.

Klappentext

This book constitutes the refereed proceedings of the 7th International Conference on Intelligent Data Analysis, IDA 2007, held in Ljubljana, Slovenia. The 33 revised papers were carefully reviewed and selected from almost 100 submissions. The book covers all current aspects of this interdisciplinary field, including statistics, machine learning, data mining, classification and pattern recognition, clustering, applications, modeling, and interactive dynamic data visualization.

Inhalt
Statistical Data Analysis.- Compact and Understandable Descriptions of Mixtures of Bernoulli Distributions.- Multiplicative Updates for L 1Regularized Linear and Logistic Regression.- Learning to Align: A Statistical Approach.- Transductive Reliability Estimation for Kernel Based Classifiers.- Bayesian Approaches.- Parameter Learning for Bayesian Networks with Strict Qualitative Influences.- Tree Augmented Naive Bayes for Regression Using Mixtures of Truncated Exponentials: Application to Higher Education Management.- Clustering Methods.- DENCLUE 2.0: Fast Clustering Based on Kernel Density Estimation.- Visualising the Cluster Structure of Data Streams.- Relational Topographic Maps.- Ensemble Learning.- Incremental Learning with Multiple Classifier Systems Using Correction Filters for Classification.- Combining Bagging and Random Subspaces to Create Better Ensembles.- Two Bagging Algorithms with Coupled Learners to Encourage Diversity.- Ranking.- Relational Algebra for Ranked Tables with Similarities: Properties and Implementation.- A New Way to Aggregate Preferences: Application to Eurovision Song Contests.- Trees.- Conditional Classification Trees Using Instrumental Variables.- Robust Tree-Based Incremental Imputation Method for Data Fusion.- Sequence/ Time Series Analysis.- Making Time: Pseudo Time-Series for the Temporal Analysis of Cross Section Data.- Recurrent Predictive Models for Sequence Segmentation.- Sequence Classification Using Statistical Pattern Recognition.- Knowledge Discovery.- Subrule Analysis and the Frequency-Confidence Diagram.- A Partial Correlation-Based Algorithm for Causal Structure Discovery with Continuous Variables.- Visualization.- Visualizing Sets of Partial Rankings.- A Partially Supervised Metric Multidimensional Scaling Algorithmfor Textual Data Visualization.- Landscape Multidimensional Scaling.- Text Mining.- A Support Vector Machine Approach to Dutch Part-of-Speech Tagging.- Towards Adaptive Web Mining: Histograms and Contexts in Text Data Clustering.- Does SVM Really Scale Up to Large Bag of Words Feature Spaces?.- Bioinformatics.- Noise Filtering and Microarray Image Reconstruction Via Chained Fouriers.- Motif Discovery Using Multi-Objective Genetic Algorithm in Biosequences.- Soft Topographic Map for Clustering and Classification of Bacteria.- Applications.- Fuzzy Logic Based Gait Classification for Hemiplegic Patients.- Traffic Sign Recognition Using Discriminative Local Features.- Novelty Detection in Patient Histories: Experiments with Measures Based on Text Compression.