node-nlp
Version:
Library for NLU (Natural Language Understanding) done in Node.js
315 lines (305 loc) • 9.35 kB
JavaScript
/*
* Copyright (c) AXA Shared Services Spain S.A.
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice shall be
* included in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
* LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
* OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
* WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*/
const { LogisticRegressionClassifier } = require('../../lib');
function getClassifier2() {
const classifier = new LogisticRegressionClassifier({});
classifier.addObservation([1, 1, 1, 0, 0, 0], 'one');
classifier.addObservation([1, 0, 1, 0, 0, 0], 'one');
classifier.addObservation([1, 1, 1, 0, 0, 0], 'one');
classifier.addObservation([0, 0, 0, 1, 1, 1], 'two');
classifier.addObservation([0, 0, 0, 1, 0, 1], 'two');
classifier.addObservation([0, 0, 0, 1, 1, 0], 'two');
return classifier;
}
function addObservations3a(classifier) {
classifier.addObservation([1, 1, 1, 0, 0, 0, 0, 0, 0], 'one');
classifier.addObservation([1, 0, 1, 0, 0, 0, 0, 0, 0], 'one');
classifier.addObservation([1, 1, 1, 0, 0, 0, 0, 0, 0], 'one');
classifier.addObservation([0, 0, 0, 1, 1, 1, 0, 0, 0], 'two');
classifier.addObservation([0, 0, 0, 1, 0, 1, 0, 0, 0], 'two');
classifier.addObservation([0, 0, 0, 1, 1, 0, 0, 0, 0], 'two');
}
function addObservations3b(classifier) {
classifier.addObservation([0, 0, 0, 0, 0, 0, 1, 1, 1], 'three');
classifier.addObservation([0, 0, 0, 0, 0, 0, 1, 0, 1], 'three');
classifier.addObservation([0, 0, 0, 0, 0, 0, 1, 1, 0], 'three');
}
function getClassifier3() {
const classifier = new LogisticRegressionClassifier({});
addObservations3a(classifier);
addObservations3b(classifier);
return classifier;
}
describe('Logistic Regression Classifier', () => {
describe('Train', () => {
test('Should create the theta', async () => {
const classifier = getClassifier2();
expect(classifier.theta).toBeUndefined();
await classifier.train();
expect(classifier.theta).toBeDefined();
const expected = [
[
2.443849977840033,
1.2755682924177316,
2.443849977840033,
-2.768423007562095,
-1.6689759372483393,
-1.6689759372483393,
],
[
-2.4438499778400344,
-1.275568292417731,
-2.4438499778400344,
2.768423007562096,
1.6689759372483401,
1.6689759372483401,
],
];
expect(classifier.theta).toHaveLength(expected.length);
for (let i = 0, l = expected.length; i < l; i += 1) {
expect(classifier.theta[i].elements).toEqual(expected[i]);
}
});
test('Should not get into an infinite loop', () => {
const classifier = new LogisticRegressionClassifier({});
classifier.addObservation([0, 0, 0, 0, 0, 0, 1, 1, 1], 'one');
classifier.addObservation([0, 0, 0, 0, 0, 0, 1, 1, 1], 'two');
});
});
describe('Get classifications', () => {
test('Should get correct clasifications for basic examples', async () => {
const classifier = getClassifier2();
await classifier.train();
const classifications1 = classifier.getClassifications([
0,
1,
1,
0,
0,
0,
]);
expect(classifications1).toHaveLength(2);
expect(classifications1[0].label).toEqual('one');
expect(classifications1[0].value).toBeGreaterThan(0.95);
expect(classifications1[1].label).toEqual('two');
expect(classifications1[1].value).toBeLessThan(0.05);
const classifications2 = classifier.getClassifications([
0,
0,
0,
0,
1,
1,
]);
expect(classifications2).toHaveLength(2);
expect(classifications2[0].label).toEqual('two');
expect(classifications2[0].value).toBeGreaterThan(0.95);
expect(classifications2[1].label).toEqual('one');
expect(classifications2[1].value).toBeLessThan(0.05);
});
test('Should get correct clasifications for more complex examples', async () => {
const classifier = getClassifier3();
await classifier.train();
const classifications1 = classifier.getClassifications([
1,
1,
0,
0,
0,
0,
1,
0,
0,
]);
expect(classifications1).toHaveLength(3);
expect(classifications1[0].label).toEqual('one');
expect(classifications1[0].value).toBeGreaterThan(0.85);
const classifications2 = classifier.getClassifications([
0,
0,
1,
1,
1,
0,
0,
0,
1,
]);
expect(classifications2).toHaveLength(3);
expect(classifications2[0].label).toEqual('two');
expect(classifications2[0].value).toBeGreaterThan(0.85);
const classifications3 = classifier.getClassifications([
1,
0,
0,
0,
1,
0,
0,
1,
1,
]);
expect(classifications3).toHaveLength(3);
expect(classifications3[0].label).toEqual('three');
expect(classifications3[0].value).toBeGreaterThan(0.6);
});
test('Should allow retraining', async () => {
const classifier = new LogisticRegressionClassifier();
addObservations3a(classifier);
await classifier.train();
let classifications1 = classifier.getClassifications([
1,
1,
0,
0,
0,
0,
1,
0,
0,
]);
expect(classifications1).toHaveLength(2);
expect(classifications1[0].label).toEqual('one');
expect(classifications1[0].value).toBeGreaterThan(0.85);
let classifications2 = classifier.getClassifications([
0,
0,
1,
1,
1,
0,
0,
0,
1,
]);
expect(classifications2).toHaveLength(2);
expect(classifications2[0].label).toEqual('two');
expect(classifications2[0].value).toBeGreaterThan(0.85);
addObservations3b(classifier);
await classifier.train();
classifications1 = classifier.getClassifications([
1,
1,
0,
0,
0,
0,
1,
0,
0,
]);
expect(classifications1).toHaveLength(3);
expect(classifications1[0].label).toEqual('one');
expect(classifications1[0].value).toBeGreaterThan(0.85);
classifications2 = classifier.getClassifications([
0,
0,
1,
1,
1,
0,
0,
0,
1,
]);
expect(classifications2).toHaveLength(3);
expect(classifications2[0].label).toEqual('two');
expect(classifications2[0].value).toBeGreaterThan(0.85);
const classifications3 = classifier.getClassifications([
1,
0,
0,
0,
1,
0,
0,
1,
1,
]);
expect(classifications3).toHaveLength(3);
expect(classifications3[0].label).toEqual('three');
expect(classifications3[0].value).toBeGreaterThan(0.6);
});
});
describe('Get Best Classification', () => {
test('Should get the best classification', async () => {
const classifier = getClassifier3();
await classifier.train();
const classification1 = classifier.getBestClassification([
1,
1,
0,
0,
0,
0,
1,
0,
0,
]);
expect(classification1.label).toEqual('one');
expect(classification1.value).toBeGreaterThan(0.85);
const classification2 = classifier.getBestClassification([
0,
0,
1,
1,
1,
0,
0,
0,
1,
]);
expect(classification2.label).toEqual('two');
expect(classification2.value).toBeGreaterThan(0.85);
const classification3 = classifier.getBestClassification([
1,
0,
0,
0,
1,
0,
0,
1,
1,
]);
expect(classification3.label).toEqual('three');
expect(classification3.value).toBeGreaterThan(0.6);
});
test('If cannot get classifications, then return undefined', async () => {
const classifier = new LogisticRegressionClassifier({});
await classifier.train();
const classification = classifier.getBestClassification([
1,
1,
0,
0,
0,
0,
1,
0,
0,
]);
expect(classification).toBeUndefined();
});
});
});