dbvis-hc

A TypeScript implementation of AGNES, an agglomerative hierarchical clustering algorithm.

Install

Install with npm:

npm install --save dbvis-hc

This package requires module resolution by Node in tsconfig.json:

{
    "compilerOptions": {
        "moduleResolution": "node"
    }   
}

Usage

This implementation can be used to cluster generic objects, given a useful definition of distance and in case of the Centroid-Linkage, additionally a definition for aggregation.

Example:

    import { CentroidLinkage, Dendrogram, HierarchicalClustering } from 'dbvis-hc';
 
    const data = [ 10, 0.9, 1.0, 11, 1.1 ];
    // Distance function required by all linkage strategies 
    const distFunc = (a: number, b: number): number => Math.abs(a - b);
    // Aggregation function only needed by the centroid linkage strategy
    const aggrFunc = (v: number[]): number => v.reduce((acc, curr) => acc + curr, 0) / v.length;
    const hc = new HierarchicalClustering<number>(data, new CentroidLinkage(distFunc, aggrFunc));
    const rootCluster = hc.cluster();
    // Split cluster according to the dendrogram and a cut-off value
    const dendrogram = new Dendrogram<number>(rootCluster, 9.3);
    const clustersAsRefs = dendrogram.extractClustersAsRefs();
    // clustersAsRefs: [ [1.1, 0.9, 1.0], [10, 11] ]
    const clustersAsIds = dendrogram.extractClustersAsIds();
    // clustersAsIds: [ [ 4, 1, 2 ], [ 3, 0 ] ]

Linkage Strategies

This package provides common linkage strategies:

• Single-Linkage
• Complete-Linkage
• Average-Linkage
• Average-Group-Linkage
• Centroid-Linkage

You can provide your own linkage method by subtyping the AbstractLinkage class.