JSPM

gradient-descent

1.0.4
  • ESM via JSPM
  • ES Module Entrypoint
  • Export Map
  • Keywords
  • License
  • Repository URL
  • TypeScript Types
  • README
  • Created
  • Published
  • Downloads 25
  • Score
    100M100P100Q71473F
  • License ISC

Package Exports

  • gradient-descent

This package does not declare an exports field, so the exports above have been automatically detected and optimized by JSPM instead. If any package subpath is missing, it is recommended to post an issue to the original package (gradient-descent) to support the "exports" field. If that is not possible, create a JSPM override to customize the exports field for this package.

Readme

Gradient Descent

Module to iterate over a numerically function to Gradient Descent direction

API

The module expose a function with next params

GD(initialPoint,asyncNumericalFunc,stepSize,deltaSize,numSteps,presicion)->Promise(bestParams)

Iterate NUM_STEPS and stop when the norm of gradient is less that PRESICION, every step is of size STEP_SIZE and the numerical derivate is aproximated by:

derivate = [f(x + DELTA_SIZE) - f(x)]/DELTA_SIZE

Usage

const optimze = require('gradient-descent')
const func = async (x, y, z) => x * x + y * y + z * z;
const init = [3, 4, 5]; // dimension is obtained from initial point
const res = await optimize(init, func);
assert(await func(...res) < await func(...init));