Title Graph classification and clustering based on vector space embedding / Kaspar Riesen & Horst Bunke

Published Singapore ; London : World Scientific, ©2010

Click on the following:
ProQuest Ebook Central
EBSCO eBooks

Copies

Description 1 online resource (xiv, 331 pages) : illustrations Series Series in machine perception and artificial intelligence ; v. 77 Series in machine perception and artificial intelligence ; v. 77. Contents 1. Introduction and basic concepts. 1.1. Pattern recognition. 1.2. Learning methodology. 1.3. Statistical and structural pattern recognition. 1.4. Dissimilarity representation for pattern recognition. 1.5. Summary and outline -- 2. Graph matching. 2.1. Graph and subgraph. 2.2. Exact graph matching. 2.3. Error-tolerant graph matching. 2.4. Summary and broader perspective -- 3. Graph edit distance. 3.1. Basic definition and properties. 3.2. Exact computation of GED. 3.3. Efficient approximation algorithms. 3.4. Exact vs. approximate graph edit distance -- an experimental evaluation. 3.5. Summary -- 4. Graph data. 4.1. Graph data sets. 4.2. Evaluation of graph edit distance. 4.3. Data visualization. 4.4. Summary -- 5. Kernel methods. 5.1. Overview and primer on kernel theory. 5.2. Kernel functions. 5.3. Feature map vs. kernel trick. 5.4. Kernel machines. 5.5. Graph kernels. 5.6. Experimental evaluation. 5.7. Summary -- 6. Graph embedding using dissimilarities. 6.1. Related work. 6.2. Graph embedding using dissimilarities. 6.3. Prototype selection strategies. 6.4. Prototype reduction schemes. 6.5. Feature selection algorithms. 6.6. Defining the reference sets for Lipschitz embeddings. 6.7. Ensemble methods. 6.8. Summary -- 7. Classification experiments with vector space embedded graphs. 7.1. Nearest-neighbor classifiers applied to vector space embedded graphs. 7.2. Support vector machines applied to vector space embedded graphs. 7.3. Summary and discussion -- 8. Clustering experiments with vector space embedded graphs. 8.1. Experimental setup and validation of the meta parameters. 8.2. Results and discussion. 8.3. Summary and discussion -- 9. Conclusions Summary This book is concerned with a fundamentally novel approach to graph-based pattern recognition based on vector space embedding of graphs. It aims at condensing the high representational power of graphs into a computationally efficient and mathematically convenient feature vector. This volume utilizes the dissimilarity space representation originally proposed by Duin and Pekalska to embed graphs in real vector spaces. Such an embedding gives one access to all algorithms developed in the past for feature vectors, which has been the predominant representation formalism in pattern recognition and related areas for a long time Bibliography Includes bibliographical references (pages 309-328) and index Notes English Print version record Subject Optical pattern recognition. Artificial intelligence -- Graphic methods Vector spaces. COMPUTERS -- Optical Data Processing. COMPUTERS -- Computer Vision & Pattern Recognition. Optical pattern recognition Vector spaces Form Electronic book Author Bunke, Horst ISBN 9789814304726 9814304727 1283144506 9781283144506 9786613144508 6613144509

  Permalink