Cluster Analysis for Applications

PrefaceAcknowledgementsChapter 1. The Broad View of Cluster Analysis 1.1 Category Sorting Problems 1.2 Need for Cluster Analysis Algorithms 1.3 Uses of Cluster Analysis 1.4 Literature of Cluster Analysis 1.5 Purpose of This BookChapter 2. Conceptual Problems in Cluster Analysis 2.1 Elements of a Cluster Analysis 2.2 Illustrative Example 2.3 Some Philosophical Observations 2.4 A Note on Optimality and IntuitionChapter 3. Variables and Scales 3.1 Classification of Variables 3.2 Scale Conversions 3.3 The Application of Scale ConversionsChapter 4. Measures of Association among Variables 4.1 Measures between Ratio and Interval Variables 4.2 Measures between Nominal Variables 4.3 Measures between Binary Variables 4.4 Strategies for Mixed Variable Data SetsChapter 5. Measures of Association among Data Units 5.1 Metric Measures for Interval Variables 5.2 Nonmetric Measures for Interval Variables 5.3 Measures Using Binary Variables 5.4 Measures Using Nominal Variables 5.5 Mixed Variable StrategiesChapter 6. Hierarchical Clustering Methods 6.1 The Central Agglomerative Procedure 6.2 The Stored Matrix Approach 6.3 The Stored Data Approach 6.4 The Sorted Matrix Approach 6.5 Other ApproachesChapter 7. Nonhierarchical Clustering Methods 7.1 Initial Configurations 7.2 Nearest Centroid Sorting-Fixed Number of Clusters 7.3 Nearest Centroid Sorting-Variable Number of Clusters 7.4 Other Approaches to Nonhierarchical ClusteringChapter 8. Promoting Interpretation of Clustering Results 8.1 Aids to Interpreting Hierarchical Classifications 8.2 An Aid to Interpreting a Partition of Data Units into ClustersChapter 9. Strategies for Using Cluster Analysis 9.1 Sequential Clustering of Data Units 9.2 Complementary Use of Several Clustering Methods 9.3 Cluster Analysis as an Adjunct to Other Statistical Methods 9.4 Clustering with Respect to an External Criterion 9.5 The Need for Research on StrategiesChapter 10. Comparative Evaluation of Cluster Analysis Methods 10.1 An Approach to the Evaluation of Clustering Methods 10.2 Quantitative Assessment of Performance for Clustering Methods 10.3 List of Candidate Characteristic for Problems and Methods 10.4 The Evaluation Task Lying AheadAppendix A. Correlation and Nominal Variables A.1 The Fundamental Analysis A.2 The Problem of Isolated Cells A.3 Deflating the Squared CorrelationAppendix B. Programs for Scale Conversions B.1 Partitions of the Truncated Normal Distribution B.2 Iterative Improvement of a Partition Program CUTS Function ERF Program DIVIDE Subroutine TEST Subroutine SORT Function PSUMSQAppendix C. Programs for Association Measures among Nominal and Interval Variables C.1 General Design Features C.2 Deck Setup and Utilization Subroutine GCORR Subroutine INPTR Subroutine NCAT Subroutine EIGEN Subroutine VSORT Function CORXX Function CORKX Function CORKKAppendix D. Programs for Association Measures Involving Binary Variables D.1 Bit-Level Storage D.2 Computing Association Measures D.3 Use of the Program Program BINARY Subroutine BDATA Function Subprogram KOUNT Function BASSNAppendix E. Programs for Hierarchical Cluster Analysis E.1 Stored Similarity Matrix Approach E.2 Stored Data Approach E.3 Sorted Matrix Approach Subroutine CNTRL Subroutine CLSTR Function LFIND Subroutine METHOD Subroutine MANAGE Subroutine GROUP Subroutine PROC Subroutine ALLINI Subroutine PREPAppendix F. Programs for Nonhierarchical Clustering Subroutine EXEC Subroutine RESULT Subroutine KMEANAppendix G. Programs to Aid Interpretation of Clustering Results G.1 A Program for Manipulating Hierarchical Trees G.2 Permuting the Similarity Matrix G.3 Error Sum of Squares Analysis G.