psl.js

JavaScript implementation of Predictive Sequence Learning.

Examples

A simple Hello world instanciation of psl.js is available here.

You may also check out the more elaborated example where psl.js is used to predict your actions in a little maze game.

Usage

psl.js can be used in the browser or standalone, using node.js.

Browser uage

Download psl.js from GitHub or linke directly to the CDN:

<script src="https://cdn.rawgit.com/billingo/psl.js/b7b7ab72/psl.js"></script>

Now, Psl can be instanciated, trained, and executed like this:

$(function() {
	var psl = new Psl(); // Instantiate the Psl library
	var s = 'PSL Hello World Hello World';
	psl.train({s:s});
	
	function run() {
		$('H1').append(psl.predict($('H1').text()));
		setTimeout(run,200);
	};

	setTimeout(run,200);
});

Please note that this example depends on jQuery. psl.js itself is however independent of external libraries. Like psl.js, jQuery can be linked directly from the CDN.

Standalone usage

First, make sure you have node.js installed. psl.js is available in the Node Package Manager (npm). Assuming that you have npm installed, simply do:

npm install psl.js

Alternatively, you may also download psl.js from GitHub and link manually.

Now, you make create a simple little Hello World like this:

var Psl = require('psl.js').Psl; // Require the Psl class from psl.js
var psl = new Psl();             // Instantiate Psl

var s = 'PSL Hello World Hello World'; // Define a source string
var h = 'P';                     // Initiate the history

process.stdout.write(h);         // Write history to console

psl.train(s);                    // Run training

function run() {
	var c = psl.predict(h);  // Predict the next character in the sequence
	process.stdout.write(c); // Print predicted character to console
	h += c;                  // Append history
	setTimeout(run,200);     // Callback
};

setTimeout(run,200);         // Initiate callback

Have fun! @billingo