Chris Ding, Hongyuan Zha

Spectral Clustering, Ordering and Ranking: Statistical Learning with Matrix Factorizations

• Gebundenes Buch

Produktbeschreibung

Data mining methods are essential for analyzing the ever-growing massive quantities of data. Data clustering is one of the key data mining techniques. In recent years, spectral clustering has been developed as an effective approach to data clustering. This exposition presents recent advances in this new subfield. New concepts are carefully developed and practical examples are extensively utilized to illustrate the ideas. A key feature are the mathematical proofs outlined throughout the text in reasonable detail which highlight the rigorous and principled quality of spectral clustering. A concise introduction to data clustering methods is followed by advanced spectral clustering, ordering and ranking topics which then lead to applications in web and text mining and genomics. An Appendix covering the preliminaries makes this text self-contained. This book is aimed at senior undergraduate and graduate students in computer science, applied mathematics and statistics and researchers and practitioners in machine learning, data mining, multivariate statistics, matrix computation, web analysis, text mining, bioinformatics.

Produktdetails

• Verlag: Springer, Berlin
• Artikelnr. des Verlages: 11561279
• 1st Edition
• Seitenzahl: 250
• Erscheinungstermin: Januar 2011
• Englisch
• Abmessung: 235mm x 155mm
• ISBN-13: 9780387304489
• ISBN-10: 0387304487
• Artikelnr.: 20742471

Inhaltsangabe

Part 1 Basic Theory and Algorithms: Introduction.- K-means clustering and spectral relaxation.- Spectral graph clustering.- Multi-way spectral clustering.- Bipartite graph clustering.- Part 2 Advanced Spectral Clustering and Ordering: Perturbation analysis of spectral clustering.- Green's function.- Random walk and probabilistic spectral clustering.- Semi-definite programming solution for spectral clustering.- Spectral data ordering.- Part 3 Related unsupervised learning topics: Spectral ranking.- Further developments on PCA and SVD.- Nonlinear embedding/representation.- Other low-dimension embedding.- Nonnegative matrix representation.- Spectral relaxations for other combinatorial optimization problems.- Level sets based clustering.- Constrained clustering.- subspace clustering.- Part 4 Applications in web, text, bioinformatics: Web and text clustering.- Clustering in genomics.- Other applications.- Appendix