This is a JavaScript reimplementation of UMAP from the python implementation found at https://github.com/lmcinnes/umap.

Uniform Manifold Approximation and Projection (UMAP) is a dimension reduction technique that can be used for visualisation similarly to t-SNE, but also for general non-linear dimension reduction.

There are a few important differences between the python implementation and the JS port.

• The optimization step is seeded with a random embedding rather than a spectral embedding. This gives comparable results for smaller datasets. The spectral embedding computation relies on efficient eigenvalue / eigenvector computations that are not easily done in JS.
• There is no specialized functionality for angular distances or sparse data representations.

yarn add umap-js

import { UMAP } from 'umap-js';

const umap = new UMAP();
const embedding = umap.fit(data);

import { UMAP } from 'umap-js';

const umap = new UMAP();
const embedding = await umap.fitAsync(data, epochNumber => {
  // check progress and give user feedback, or return `false` to stop
});

import { UMAP } from 'umap-js';

const umap = new UMAP();
const nEpochs = umap.initializeFit(data);
for (let i = 0; i < nEpochs; i++) {
  umap.step();
}
const embedding = umap.getEmbedding();

import { UMAP } from 'umap-js';

const umap = new UMAP();
umap.setSupervisedProjection(labels);
const embedding = umap.fit(data);

import { UMAP } from 'umap-js';

const umap = new UMAP();
umap.fit(data);
const transformed = umap.transform(additionalData);

The UMAP constructor can accept a number of hyperparameters via a UMAPParameters object, with the most common described below. See umap.ts for more details.

const umap = new UMAP({
  nComponents: 2,
  nEpochs: 400,
  nNeighbors: 15,
});

umap-js uses jest for testing.

yarn test

This is not an officially supported Google product