UNPKG

deriv

Version:

numerical derivatives of one-variable functions

36 lines (25 loc) 1.35 kB
[![NPM version](https://badge.fury.io/js/deriv.svg)](http://badge.fury.io/js/deriv) [![Build Status](https://travis-ci.org/scijs/deriv.svg?branch=master)](https://travis-ci.org/scijs/deriv) deriv ========== numerical derivatives of one-variable functions # Installation and Usage Installation via npm: ``` npm install deriv ``` Requiring the function takes one line: ``` var deriv = require("deriv"); ``` Evaluating a derivative also: ``` x_prime = deriv(function(x){return x*x}, 3, 0.1); console.log(x_prime) ``` Output: ``` { res: 5.999999999999166, err: 0.15000000000437436 } ``` # deriv(fun, x, h, [type]) The first parameter is the function for which the derivative should be evaluated. The function should have only one parameter, expecting a numeric value. The second parameter `x` is the value at which the derivative should be calculated, whereas `h` is an appropriately chosen step size. `type` is an optional string parameter, which can take either the value `forward` or `central`. Specifically, this module implements the algorithms used in the numerical differentiation functions of the [GSL library](https://www.gnu.org/software/gsl/manual/html_node/Numerical-Differentiation-functions.html) and described in the book "S.D. Conte and Carl de Boor, Elementary Numerical Analysis: An Algorithmic Approach, McGraw-Hill, 1972".