TF-KMeans

Description

A Simple JavaScript Library to make it easy for people to use KMeans algorithms with Tensorflow JS.

The library was born out of another project in which except KMeans, our code completely depended on TF.JS

As such, moving to TF.JS helped standardise our code base substantially and reduce dependency on other libraries

Sample Code

    const KMeans = require("tf-kmeans");
    const tf = require("@tensorflow/tfjs");
    const kmeans = new KMeans.default({
        k: 2,
        maxIter: 30,
        distanceFunction: KMeans.default.EuclideanDistance
    });
    const dataset = tf.tensor([[2, 2, 2], [5, 5, 5], [3, 3, 3], [4, 4, 4], [7, 8, 7]]);
    const predictions = kmeans.Train(
        dataset
    );
 
    console.log("Assigned To ", predictions.arraySync());
    console.log("Centroids Used are ", kmeans.Centroids().arraySync());
    console.log("Prediction for Given Value is");
    kmeans.Predict(tf.tensor([2, 3, 2])).print();

You can use the Asynchronous TrainAsync if you want to use an asynchronous callback function

    const kmeans = new KMeans.default({
        k: 3,
        maxIter: 30,
        distanceFunction: KMeans.default.EuclideanDistance
    });
    const dataset = tf.tensor([[2, 2, 2], [5, 5, 5], [3, 3, 3], [4, 4, 4], [7, 8, 7]]);
 
    console.log("\n\nAsync Test");
    const predictions = await kmeans.TrainAsync(
        dataset,
        // Called At End of Every Iteration
        // This function is Asynchronous
        async(iter, centroid, preds)=>{
            console.log("===");
            console.log("Iteration Count", iter);
            console.log("Centroid ", await centroid.array());
            console.log("Prediction ", await preds.array());
            console.log("===");
            // You could instead use TFVIS for Plotting Here
        }
    );

Functions

1. Constructor

   Takes 3 Optional parameters

   1. k:- Number of Clusters
   2. maxIter:- Max Iterations
   3. distanceFunction:- The Distance function Used Currently only Eucledian Distance Provided

2. Train

   Takes Dataset as Parameter

   Performs Training on This Dataset

   Sync callback function is optional

3. TrainAsync

   Takes Dataset as Parameter

   Performs Training on This Dataset

   Also takes async callback function called at the end of every iteration

4. Centroids

   Returns the Centroids found for the dataset on which KMeans was Trained

5. Predict

   Performs Predictions on the data Provided as Input

PEER DEPENDENCIES

1. TensorFlow.JS

Typings

As the code is originally written in TypeScript, Type Support is provided out of the box

Contact Me

You could contact me via LinkedIn You could file issues or add features via Pull Requests on GitHub