Base class for regression modules

Usage no npm install needed!

<script type="module">
  import mlRegressionBase from '';



NPM version build status Test coverage npm download

Base class for regression modules.
This package is for ml.js internal use.


You only have to implement the _predict method. It is always called with a number.

The model should be created in the constructor.

Optional methods that can be implemented: toString, toLaTeX.

import BaseRegression from 'ml-regression-base';

class MyRegression extends BaseRegression {
  constructor(factor) {
    this.factor = factor;
  _predict(x) {
    return x * this.factor;
  toString() {
    return `f(x) = x * ${this.factor}`;

maybeToPrecision(value, digits)

Convenience method to transform numbers to readable strings.

If digits is not specified, "value.toString()" is used. Otherwise "value.toPrecision(digits)" is used.

This method can be used to implement toString() or toLaTeX().

checkArrayLength(x, y)

Convenience method to check if the input and output arrays passed to a regression constructor are effectively arrays with the same length.