Batagelj / Bock / Ferligoj Data Science and Classification

1. Auflage 2006
ISBN: 978-3-540-34416-2
Verlag: Springer-Verlag
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)

E-Book, Englisch, 358 Seiten

Data Science and Classification provides new methodological developments in data analysis and classification. The broad and comprehensive coverage includes the measurement of similarity and dissimilarity, methods for classification and clustering, network and graph analyses, analysis of symbolic data, and web mining. Beyond structural and theoretical results, the book offers application advice for a variety of problems, in medicine, microarray analysis, social network structures, and music.

Autoren/Hrsg.

Batagelj, Vladimir

Bock, Hans-Hermann

Ferligoj, Anuška

Žiberna, Aleš

Inhaltsverzeichnis

1;Preface;6
2;The 10th IFCS Conference – a Jubilee;7
3;Contents;9
4;Similarity and Dissimilarity;13
4.1;A Tree-Based Similarity for Evaluating Concept Proximities in an Ontology;14
4.2;Improved Frechet Distance for Time Series;23
4.3;Comparison of Distance Indices Between Partitions;31
4.4;Design of Dissimilarity Measures: A New Dissimilarity Between Species Distribution Areas;39
4.5;Dissimilarities for Web Usage Mining;48
4.6;Properties and Performance of Shape Similarity Measures;56
5;Classification and Clustering;66
5.1;Hierarchical Clustering for Boxplot Variables;67
5.2;Evaluation of Allocation Rules Under Some Cost Constraints;75
5.3;Crisp Partitions Induced by a Fuzzy Set;82
5.4;Empirical Comparison of a Monothetic Divisive Clustering Method with the Ward and the k- means Clustering Methods;90
5.5;Model Selection for the Binary Latent Class Model: A Monte Carlo Simulation;98
5.6;Finding Meaningful and Stable Clusters Using Local Cluster Analysis;107
5.7;Comparing Optimal Individual and Collective Assessment Procedures;115
6;Network and Graph Analysis;123
6.1;Some Open Problem Sets for Generalized Blockmodeling;124
6.2;Spectral Clustering and Multidimensional Scaling: A Unified View;136
6.3;Analyzing the Structure of U.S. Patents Network;145
6.4;Identifying and Classifying Social Groups: A Machine Learning Approach;153
7;Analysis of Symbolic Data;162
7.1;Multidimensional Scaling of Histogram Dissimilarities;163
7.2;Dependence and Interdependence Analysis for Interval- Valued Variables;173
7.3;A New Wasserstein Based Distance for the Hierarchical Clustering of Histogram Symbolic Data;186
7.4;Symbolic Clustering of Large Datasets;194
7.5;A Dynamic Clustering Method for Mixed Feature- Type Symbolic Data;203
8;General Data Analysis Methods;211
8.1;Iterated Boosting for Outlier Detection;212
8.2;Sub-species of Homopus Areolatus? Biplots and Small Class Inference with Analysis of Distance;220
8.3;Revised Boxplot Based Discretization as the Kernel of Automatic Interpretation of Classes Using Numerical Variables;228
9;Data and Web Mining;237
9.1;Comparison of Two Methods for Detecting and Correcting Systematic Error in High- throughput Screening Data;238
9.2;kNN Versus SVM in the Collaborative Filtering Framework;247
9.3;Mining Association Rules in Folksonomies;257
9.4;Empirical Analysis of Attribute-Aware Recommendation Algorithms with Variable Synthetic Data;267
9.5;Patterns of Associations in Finite Sets of Items;275
10;Analysis of Music Data;283
10.1;Generalized N-gram Measures for Melodic Similarity;284
10.2;Evaluating Different Approaches to Measuring the Similarity of Melodies;294
10.3;Using MCMC as a Stochastic Optimization Procedure for Musical Time Series;302
10.4;Local Models in Register Classification by Timbre;310
11;Gene and Microarray Analysis;318
11.1;Improving the Performance of Principal Components for Classification of Gene Expression Data Through Feature Selection;319
11.2;A New Efficient Method for Assessing Missing Nucleotides in DNA Sequences in the Framework of a Generic Evolutionary Model;327
11.3;New Efficient Algorithm for Modeling Partial and Complete Gene Transfer Scenarios;335
12;List of Reviewers;344
13;Key words;346
14;Authors;349