linearreg - reglin/linearreg.ts - Ubique documentation

function linearreg



linearreg(y: array,
x: array,
): { beta: number; alpha: number; rsq: number; fun: (x: number) => number;  }

Linear regression of Y on X.

Performs linear regression and returns an object with regression coefficients, R-squared value, and a prediction function.

Examples

Basic linear regression

import { assertEquals } from "jsr:@std/assert";

const x = [0.003,0.026,0.015,-0.009,0.014,0.024,0.015,0.066,-0.014,0.039];
const y = [-0.005,0.081,0.04,-0.037,-0.061,0.058,-0.049,-0.021,0.062,0.058];
const result = linearreg(y,x);
assertEquals(Math.round(result.beta * 1000) / 1000, 0.119);
assertEquals(Math.round(result.alpha * 1000) / 1000, 0.01);

Linear regression with stronger correlation

import { assertEquals } from "jsr:@std/assert";

const result2 = linearreg([100,101,99,102,105],[1,2,3,4,5]);
assertEquals(result2.beta, 1.1);
assertEquals(result2.alpha, 98.1);

Using the prediction function

import { assertEquals } from "jsr:@std/assert";

const result2 = linearreg([100,101,99,102,105],[1,2,3,4,5]);

assertEquals(Math.round(result2.fun(6) * 10) / 10, 104.7);

Parameters



y: array

array of dependent variable values



x: array

array of independent variable values

Return Type

{ beta: number; alpha: number; rsq: number; fun: (x: number) => number;  }

Object with beta (slope), alpha (intercept), rsq (R-squared), and prediction function