Chi-Square

This package contains functions to calculate Chi-Square critical values as well as performing a Chi-Square statistical test.

Installation

npm install @folge3/chi-square

Usage

import {chiSquare, DistributionData, Distribution, ChiSquareTestResult, chiSquareCriticalValue, ChiSquareOptions} from "@folge3/chi-square";

const data: DistributionData = {
    rows: ['row1', 'row2'],
    columns: ['column1', 'column2'],
    data: [
        [52, 27],
        [23, 56]
    ]
}

const distribution: Distribution = new Distribution(data);

const options: ChiSquareOptions = {
    rowImpactWeightExponent: 0.3
}

// Run chi-square test with 5% uncertainty 
const result: ChiSquareTestResult = chiSquare(distribution, 0.05, options);

console.log(result);

// Get critical value for significance level 0.05 and 1 degree of freedom. 
const criticalValue = chiSquareCriticalValue(0.05, 1);

// ~3.841
console.log(criticalValue)

Options

Here are the available options for the ChiSquareOptions:

|Option| Type | Description | |------|--------|----------------------------------------------------------------------------------------------| |rowImpactWeightExponent| number | Exponent modifier for weights in row impact rating. Must lie in [0.1, 1]. Default: 1. | |normalizeRows| boolean | Flag to decide if data shall be normalized by its rows. Takes precedence over columns. | |normalizeColumns| boolean | Flag to decide if data shall be normalized by its columns. Is preceded by row normalization. | |normalizationMarginalTarget| number | Target value for marginals for normalization. Must be greater than 0. Default: 100. |

Development

Testing

npm run test

Deployment

1. Commit changes
2. npm version [patch|minor|major]
3. npm run build
4. npm publish --access public