simple-clustering

Finds the isolated clusters from a similarity/connectivity matrix.

The input could be a matrix (array of arrays) or it could be a single array containing the values of the upper triangular of that matrix.

By default the function returns a clustering matrix: It is a binary row wise matrix, where each row has n components. The ones at row i and column j indicates that element j belong to cluster i.

Installation

$ npm install ml-simple-clustering

Usage

import { simpleClustering } from 'ml-simple-clustering';

const clusters = simpleClustering(dataMatrix, options);
Copy

API Documentation

In order to get a general idea of the problem you could also check the Wikipedia article.

Examples

const { simpleClustering } = require('ml-simple-clustering');

let matrix = [
  [1, 2, 0, 0],
  [1, 2, 0, 0],
  [0, 0, 3, 4],
  [0, 0, 5, 6],
];

let clusters = simpleClustering(matrix, {
  threshold: 0,
  out: 'values',
});
console.log(`clusters = ${clusters}`);
/**
 clusters = [
  [
    [1, 2],
    [1, 2],
  ],
  [
    [3, 4],
    [5, 6],
  ],
]
  */
Copy

License

MIT