Illustrations of Key Clustering Concepts

Document Space, Clusters, Centroids

• Document Collection as A Space with Points
• Clustering of A Space, Cluster Centroids

Low Level Centroids, Hierarchical Clustering, Cluster Search

• Low Level Centroids, Hierarchical Clustering
• Search of LLC or Cluster Hierarchy

Agglomerative Clustering, Dendogram

• Space Partitioned into 11 Regions
• Cluster Step 1: Join Regions 8-9
• Cluster Step 2: Join Regions 1-3
• Cluster Step 3: Join Regions 6-7
• Cluster Step 4: Join Regions 8-9 to 11
• Rest of Clustering Example as a Dendogram

Document-Document Similarity Matrix, MST, NNs

• Document Similarity Matrix
• Sim. Matrix - Minimal Spanning Tree
• Sim. Matrix - Nearest Neighbors

Test of Clustering Tendency