npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

node-ml

v1.0.2

Published

A Collection of Machine Learning algorithms built for use with NodeJS

Downloads

15

Readme

node-ml

( Development is current suspended ).

A Collection of Machine Learning algorithms built for use with NodeJS

build status

Installation

npm install node-ml

Basic API Knowledge

Models are Instantiated with a TrainingSet as an argument

Models are trained via .train()

Models are acted upon by the perceive or predict functions. These functions currently do the same thing, the wording is indicative of the nature of the result, and the action the model has taken on the data.

Models are designed to be trained at the start of a node instance, not during runtime.

Models all inherit from EventEmitter

Models support either Callbacks or Events

callbacks are of the structure function( err, result )

Models support 3 events : trained, response, error

Callbacks override Events, if you specify a callback and event will not be received.

The Single Layer Perceptron

With the single layer perceptron is it possible to solve Linearly Seperable Problems. This makes the SLP a fast tool for solving simple classification problems.

The SLP takes as input a list of 1x2 vectors as in

[
  [1,1],
  [-1,-1]
]

We must also provide the SLP a list of expected outputs for each vector. Currently the system only supports 1 | -1 These outputs define the side of the line the elements fall on. It is not important which value you give to which inputs. Just that these inputs correspond in a linear way to the outputs.

So for the above input we prove

[
  1,
  -1
]

Now the SLP will solve for where [1,1] === 1 and [-1,-1] === -1

Training the SLP

SingleLayerPerceptron(inputs, outputs, learningRate)

slp = new SingleLayerPerceptron(inputs, outputs, 0.001);
slp.train(function(trainedModel) { 
    trainedModel.perceive([1,1], function(result) {
      console.log(result);
      //should print out 1
    }); 
    trainedModel.perceive([-1,-1], function(result) {
      console.log(result);
      //should print out -1
    });
});

Also Events are supported

current events are: trained, response, error

trained is fired when a model completes training

response is fired when a model completed a perception or prediction phase

error is fired ... well on an error.

    slp = new SingleLayerPerceptron(inputs, outputs, 0.001);
    slp.on('error', function(err) {
        t.same(1,1);
    });
    slp.on('trained', function(trainedModel) {
        t.same(true,(trainedModel != undefined));

        trainedModel.perceive([1,1]);
        trainedModel.perceive([-1,-1]);
    });
    slp.on('response', function(response) {
        perceivedTestCount -= 1;
        var result = response.out;
        var input = response.in;
        var expectedIndex;
        for (var i in perceivedTestInput ) {
            if (perceivedTestInput[i].toString() == input.toString())
                expectedIndex = i;
            
        }
        var expected = perceivedTestOutput[expectedIndex];
        t.same(expected, result);

        if (perceivedTestCount == 0 )
            t.end();
        
    });

    slp.train();

Even Better remember the above trained model is a Line seperating a 2d dimension space from -1 to 1 We can input any value in this range and get an output however this limited training set is a bad choice but heres some outputs

Obtained From Running examples/singlelayerperceptron2.js

Input: 1,1
1
Input: -1,1
-1
Input: 1,-1
1
Input: -.5,1
1
Input: .5,-1
-1
Input: .2,.45634
1
Input: .2,-.45634
-1
Input: -.4,-.4
-1
Input: -1,-1
-1

The Multi Layer Perceptron

With the Multi Layer Perceptron it is possible to Classify linearly non seperable data set. Meaning that the data fits to a polynomial function.

Refer to examples.

The Linear Regression Model

With Linear Regression we can predict outcomes based on an input.

Refer to examples.

The KMeans Classifier

This implementation of the KMeans classifier is an N-Dimensional classification algorithm. It works by:

  1. Generating Random K Centroids.

  2. Assigning a K centroid to a point p in Training set T such that the Cost(p) with respect to k is minimized; Cost(p) is defined as the distances from P to each K

  3. Move the centroids to the Mean of each cluster assigned to them

  4. Repeat until movement no longer occurs.

Refer to examples for Usage:

Bitdeli Badge