@jsmlt/jsmlt/distribution/kernel/gaussian.test.js

JavaScript Machine Learning

"use strict";

var expect = require('chai').expect;

var GaussianKernel = require('./gaussian.js');

describe('Kernel.GaussianKernel', function () {
  describe('.apply', function () {
    it('should calculate the Gaussian kernel with variance 1 when called with default parameters', function () {
      expect(new GaussianKernel().apply([1, 2], [3, 4])).to.closeTo(0.01831563888, 1e-5);
    });
    it('should correctly calculate the Gaussian kernel with custom variance', function () {
      expect(new GaussianKernel(2).apply([1, 2], [3, 4])).to.closeTo(0.13533528323, 1e-5);
    });
    it('should return 1 when a point is compared to itself', function () {
      expect(new GaussianKernel().apply([1, 2], [1, 2])).to.closeTo(1, 1e-5);
    });
  });
});