Crate catclustering

CatClustering

catclustering implements hierarchical agglomerative clustering optimized for categorical features and (approximated) complete-linkage. Explanation and examples here.

Traits

ClusterSummary

The trait you need to implement to provide the algorithm a distance and merging strategy.

IndexableData

The trait you need to implement for the clustering algorithm to access your data.

Functions

assign_rows_to_clusters

This assigns a cluster number to each row under the constraint that all clusters must have the same size or be smaller than the provided size.

create_dendrogram

This function groups input data into clusters and returns the corresponding dendrogram.

find_clusters

This traverses the dendrogram until it finds clusters of the same size or smaller than the given size, and returns these clusters.