Random Graphs for Statistical Pattern Recognition (eBook)

DAVID J. MARCHETTE, PhD, is a researcher at the Naval Surface Warfare Center in Dahlgren, Virginia, where he investigates computational statistics and pattern recognition, primarily as it applies to image processing, automatic target recognition, and computer security. He is also an adjunct professor at George Mason University and a lecturer at Johns Hopkins University.

Preface.

Acknowledgments.

1. Preliminaries.

1.1 Graphs and Digraphs.

1.2 Statistical Pattern Recognition.

1.3 Statistical Issues.

1.4 Applications.

1.5 Further Reading.

2. Computational Geometry.

2.1 Introduction.

2.2 Voronoi Cells and Delaunay Triangularization.

2.3 Alpha Hulls.

2.4 Minimum Spanning Trees.

2.5 Further Reading.

3. Neighborhood Graphs.

3.1 Introduction.

3.2 Nearest-Neighbor Graphs.

3.3 k-Nearest Neighbor Graphs.

3.4 Relative Neighborhood Graphs.

3.5 Gabriel Graphs.

3.6 Application: Nearest Neighbor Prototypes.

3.7 Sphere of Influence Graphs.

3.8 Other Relatives.

3.9 Asymptotics.

3.10 Further Reading.

4. Class Cover Catch Digraphs.

4.1 Catch Digraphs.

4.2 Class Covers.

4.3 Dominating Sets.

4.4 Distributional Results for Cn,m-graphs.

4.5 Characterizations.

4.6 Scale Dimension.

4.7 (alpha,beta) Graphs

4.8 CCCD Classification.

4.9 Homogeneous CCCDs.

4.10 Vector Quantization.

4.11 Random Walk Version.

4.12 Further Reading.

5. Cluster Catch Digraphs.

5.1 Basic Definitions.

5.2 Dominating Sets.

5.3 Connected Components.

5.4 Variable Metric Clustering.

6. Computational Methods.

6.1 Introduction.

6.2 Kd-Trees.

6.3 Class Cover Catch Digraphs.

6.4 Cluster Catch Digraphs.

6.5 Voroni Regions and Delaunay Triangularizations.

6.6 Further Reading.

References.

Author Index.

Subject Index.

"...constructed...as a book on random graphs, this is quite
a good one." (Journal of the American Statistical
Association, September 2006)

"...I recommend this book to those who...wish to explore the
exciting place where graph theory and pattern recognition meet."
(Statistics in Medical Research, October 2005)

"This well-written book presents practical tools, and
information that was previously found scattered in various
journals." (Computing Reviews.com, March 9, 2005)

"...an excellent resource book that would be a valuable
addition..." (Technometrics, February 2005)

"...clearly and accessible written, and nicely conveys
the power, breadth and applicability of some very elegant
ideas..." (Short Book Reviews, Vol.24, No.3, December
2004)

"Buy this book if use graphs in cluster and classification
analysis." (Journal of Classification, Vol.21, No.2,
2004)