Examining the Validity of Sample Clusters, Using the Bootstrap Method (Classic Reprint) - Hardcover

M. A. Wong

9780666362803: Examining the Validity of Sample Clusters, Using the Bootstrap Method (Classic Reprint)

Hardcover

ISBN 10: 0666362807 ISBN 13: 9780666362803

Publisher: Forgotten Books, 2018

View all copies of this ISBN edition

0 Used

2 New

From US$ 38.95

Explore how bootstrap analysis tests cluster stability in data analysis and what that means for real results.

This book from Examining the Validity of Sample Clusters, Using the Bootstrap Method focuses on using Bk plots to judge when clusters are stable. It compares how different linkage methods and data set configurations affect the appearance of stable groups, and it discusses the value of bootstrapping in understanding variability.

The material walks through practical insights on reading trees and Bk versus k plots, noting how peaks can indicate robust groupings and how results change with complete, single, and average linkage. It also covers how bootstrapped estimates compare to the original samples and what that implies for interpreting clustering results in real data.

How to read Bk plots and identify likely stable clusters across methods and data sets.
how different linkage approaches (complete, single, average) shape cluster stability and peak behavior.
The role of bootstrap samples in estimating the true distribution of Bk and the value of median, mean, and mode as estimators.
Practical cautions on interpreting peaks, null distributions, and the limits of bootstrap-based inferences.

Ideal for readers of clustering analysis and bootstrap techniques, this edition offers concrete examples and guidance for evaluating when clusters really hold up under resampling.

"synopsis" may belong to another edition of this title.