discrete-sampling

sampling

JavaScript for sampling from a handful of discrete probability distributions

Setup

This is now a node module on npm. To install, go to your project's directory in a terminal and enter npm install discrete-sampling To load the library in your own node module, use

var SJS = Sampling = require('discrete-sampling');

To load the module in the browser, you can load the script with

<script src="node_modules/discrete-sampling/dist/discrete.js"></script>

and then the library will be accessible through the global variables SJS and Sampling.

Currently Supported

• Bernoulli
• Binomial
• Discrete
• Multinomial
• Poisson
• Negative Binomial

Examples

var bern = Sampling.Bernoulli(.5);
bern.draw();     // One random draw
bern.sample(10); // Get an array of 10 samples. 

var binom = SJS.Binomial(10, 0.5);  // alternate syntax (Sampling === SJS)
binom.draw();
binom.sample(10);

var probabilities = [.1, .2, .3, .05, .15];
var disc = SJS.Discrete(probabilities);
disc.draw(); 
disc.sample(10); 

// The probabilities will be normalized for you. 
var probs = [1,1,1,1,1,1,1,1,1,1,1,1,1,1]
var mult = SJS.Multinomial(5, probs);
mult.draw();
mult.sample(10);

var myArray = [0,1,2,3,4,5]
// sample five elements with replacement
sample_from_array(myArray, 5, true)
// sample five elements without replacement
sample_from_array(myArray, 5)