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

encog

v1.6.0

Published

Encog is a NodeJs ES6 framework based on the Encog Machine Learning Framework by Jeff Heaton, plus some the of basic data manipulation helpers.

Downloads

126

Readme

encog

https://www.npmjs.com/package/encog

Encog is a NodeJs ES6 framework based on the Encog Machine Learning Framework by Jeff Heaton.

All credits of the framework should go to Jeff Heaton - http://www.heatonresearch.com/encog/

Based on the encog-java-core v3.4 - https://github.com/encog/encog-java-core

Full documentation and source code - https://github.com/redsoul/encog

Build Status

Installation

npm install encog --save

Usage

Just require the library and all of Encog namespace will be available to you:

const Encog = require('encog');

Unit Tests

npm install --only=dev
npm test

Implemented algorithms

  • Networks
    • Basic Network
    • Hopfield Network
    • BAM (Bidirectional associative memory) Network
    • Freeform Network
  • Training
    • Back Propagation
    • Manhattan Propagation
    • Resilient Propagation
    • Stochastic Gradient Descent
      • Momentum
      • Nesterov
      • RMS Prop
      • AdaGrad
      • Adam
    • Levenberg Marquardt
    • Neural Simulated Annealing
  • Patterns
    • ADALINE
    • Feed Forward (Perceptron)
    • Elman Network
    • Jordan Network
    • Hopfield Network
    • BAM Network
  • Activation Functions
    • Elliott
    • Symmetric Elliott
    • Gaussian
    • Linear
    • Ramp
    • ReLu
    • Sigmoid
    • Softmax
    • Steepened Sigmoid
    • Hyperbolic tangent
  • Error Functions
    • Arctangent
    • Cross Entropy
    • Linear
    • Output

Examples

Back Propagation example using XOR Data Set

const Encog = require('encog');
const XORdataset = Encog.Utils.Datasets.getXORDataSet();

//adjust the log level
Encog.Log.options.logLevel = 'info';

// create a neural network
const network = new Encog.Networks.Basic();
network.addLayer(new Encog.Layers.Basic(null, true, 2));
network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), true, 4));
network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), false, 1));
network.randomize();

const train = new Encog.Training.Propagation.Back(network, XORdataset.input, XORdataset.output);

Encog.Utils.Network.trainNetwork(train, {maxIterations: 250});
const accuracy = Encog.Utils.Network.validateNetwork(network, XORdataset.input, XORdataset.output);
console.log('Accuracy:', accuracy);

Resilient Propagation example using Iris Flower Data Set (https://en.wikipedia.org/wiki/Iris_flower_data_set)

const Encog = require('encog');
const _ = require('lodash');

//adjust the log level
Encog.Log.options.logLevel = 'info';

const dataEncoder = new Encog.Preprocessing.DataEncoder();
let irisDataset = Encog.Utils.Datasets.getIrisDataSet();
irisDataset = _.shuffle(irisDataset);
irisDataset = Encog.Preprocessing.DataToolbox.trainTestSplit(irisDataset);

/******************/
//data normalization
/******************/

//apply a specific mapping to each column
const mappings = {
    'Sepal.Length': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
    'Sepal.Width': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
    'Petal.Length': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
    'Petal.Width': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
    'Species': new Encog.Preprocessing.DataMappers.OneHot(),
};

//Fit to data, then transform it.
let trainData = dataEncoder.fit_transform(irisDataset.train, mappings);
//transform the test data based on the train data
let testData = dataEncoder.transform(irisDataset.test, mappings);

//slice the data in input and output
trainData = Encog.Preprocessing.DataToolbox.sliceOutput(trainData.values, 3);
testData = Encog.Preprocessing.DataToolbox.sliceOutput(testData.values, 3);

// create a neural network
const network = new Encog.Networks.Basic();
network.addLayer(new Encog.Layers.Basic(null, true, 4));
network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), true, 10));
network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), true, 5));
network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), false, 3));
network.randomize();

// train the neural network
const train = new Encog.Training.Propagation.Resilient(network, trainData.input, trainData.output);
Encog.Utils.Network.trainNetwork(train, {minError: 0.01, minIterations: 5});

//validate the neural network
let accuracy = Encog.Utils.Network.validateNetwork(network, testData.input, testData.output);
console.log('Accuracy:', accuracy);

//save the trained network
Encog.Utils.File.saveNetwork(network, 'iris.dat');

//load a pretrained network
const newNetwork = Encog.Utils.File.loadNetwork('iris.dat');

//validate the neural network
accuracy = Encog.Utils.Network.validateNetwork(newNetwork, testData.input, testData.output);
console.log('accuracy: ', accuracy);

Stochastic Gradient Descent with Adam update example using the bank note authentication dataset

const Encog = require('encog');
const _ = require('lodash');
const dataEncoder = new Encog.Preprocessing.DataEncoder();

//adjust the log level
Encog.Log.options.logLevel = 'info';

(async () => {
    const dataset = await Encog.Preprocessing.DataToolbox.readTrainingCSV(
        './node_modules/encog/examples/data/data_banknote_authentication.csv'
    );
    const shuffledDataset = _.shuffle(dataset);

    const splittedDataset = Encog.Preprocessing.DataToolbox.trainTestSplit(shuffledDataset);

    /******************/
    //data normalization
    /******************/
    //apply a specific mapping to each column
    const mappings = {
        'variance': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
        'skewness': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
        'curtosis': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
        'entropy': new Encog.Preprocessing.DataMappers.MinMaxScaller(),
        'class': new Encog.Preprocessing.DataMappers.IntegerParser()
    };
    //Fit to data, then transform it.
    let trainData = dataEncoder.fit_transform(splittedDataset.train, mappings);
    //transform the test data based on the train data
    let testData = dataEncoder.transform(splittedDataset.test, mappings);

    //slice the data in input and output
    trainData = Encog.Preprocessing.DataToolbox.sliceOutput(trainData.values, 1);
    testData = Encog.Preprocessing.DataToolbox.sliceOutput(testData.values, 1);

    // create a neural network
    const network = new Encog.Networks.Basic();
    network.addLayer(new Encog.Layers.Basic(null, true, 4));
    network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), true, 40));
    network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), true, 40));
    network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), true, 40));
    network.addLayer(new Encog.Layers.Basic(new Encog.ActivationFunctions.Sigmoid(), false, 1));
    network.randomize();

    // train the neural network
    const train = new Encog.Training.SGD.StochasticGradientDescent(network, trainData.input, trainData.output, new Encog.Training.SGD.Update.Adam());
    Encog.Utils.Network.trainNetwork(train, {minError: 0.01, minIterations: 50, maxIterations: 200});

    //validate the neural network
    let accuracy = Encog.Utils.Network.validateNetwork(network, testData.input, testData.output);
    console.log('Accuracy:', accuracy);

    //save the trained network
    Encog.Utils.File.saveNetwork(network, 'banknote_authentication.dat');

    //load a pretrained network
    const newNetwork = Encog.Utils.File.loadNetwork('banknote_authentication.dat');

    //validate the neural network
    accuracy = Encog.Utils.Network.validateNetwork(newNetwork, testData.input, testData.output);
    console.log('accuracy: ', accuracy);
})();

Hopfield Network example custom binary dataset

const Encog = require('encog');
const _ = require('lodash');
const hopfieldPatterns = Encog.Utils.Datasets.getHopfieldPatterns();
const HopfieldPattern = new Encog.Patterns.Hopfield();

//adjust the log level
Encog.Log.options.logLevel = 'info';

HopfieldPattern.setInputLayer(35);
const network = HopfieldPattern.generate();

_.each(hopfieldPatterns, function (pattern) {
    network.addPattern(pattern);
});

network.runUntilStable(10);
const input = [
    0, 0, 0, 0, 0,
    0, 1, 1, 1, 0,
    0, 0, 0, 0, 0,
    0, 1, 1, 0, 0,
    0, 0, 0, 0, 0,
    0, 1, 1, 1, 0,
    0, 0, 0, 0, 0
];
const result = network.compute(input);
console.log('Result:', result);

/*
Output:

0, 0, 0, 0, 0,
0, 1, 1, 1, 0,
0, 1, 0, 0, 0,
0, 1, 1, 0, 0,
0, 1, 0, 0, 0,
0, 1, 1, 1, 0,
0, 0, 0, 0, 0
*/

Node.js version compatibility

8.0.0 or higher