122 lines
3.6 KiB
JavaScript
122 lines
3.6 KiB
JavaScript
/**
|
||
* @license Apache-2.0
|
||
*
|
||
* Copyright (c) 2020 The Stdlib Authors.
|
||
*
|
||
* Licensed under the Apache License, Version 2.0 (the "License");
|
||
* you may not use this file except in compliance with the License.
|
||
* You may obtain a copy of the License at
|
||
*
|
||
* http://www.apache.org/licenses/LICENSE-2.0
|
||
*
|
||
* Unless required by applicable law or agreed to in writing, software
|
||
* distributed under the License is distributed on an "AS IS" BASIS,
|
||
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||
* See the License for the specific language governing permissions and
|
||
* limitations under the License.
|
||
*/
|
||
|
||
'use strict';
|
||
|
||
// MODULES //
|
||
|
||
var float64ToFloat32 = require( '@stdlib/number/float64/base/to-float32' );
|
||
|
||
|
||
// MAIN //
|
||
|
||
/**
|
||
* Computes the variance of a single-precision floating-point strided array using Welford's algorithm.
|
||
*
|
||
* ## Method
|
||
*
|
||
* - This implementation uses Welford's algorithm for efficient computation, which can be derived as follows. Let
|
||
*
|
||
* ```tex
|
||
* \begin{align*}
|
||
* S_n &= n \sigma_n^2 \\
|
||
* &= \sum_{i=1}^{n} (x_i - \mu_n)^2 \\
|
||
* &= \biggl(\sum_{i=1}^{n} x_i^2 \biggr) - n\mu_n^2
|
||
* \end{align*}
|
||
* ```
|
||
*
|
||
* Accordingly,
|
||
*
|
||
* ```tex
|
||
* \begin{align*}
|
||
* S_n - S_{n-1} &= \sum_{i=1}^{n} x_i^2 - n\mu_n^2 - \sum_{i=1}^{n-1} x_i^2 + (n-1)\mu_{n-1}^2 \\
|
||
* &= x_n^2 - n\mu_n^2 + (n-1)\mu_{n-1}^2 \\
|
||
* &= x_n^2 - \mu_{n-1}^2 + n(\mu_{n-1}^2 - \mu_n^2) \\
|
||
* &= x_n^2 - \mu_{n-1}^2 + n(\mu_{n-1} - \mu_n)(\mu_{n-1} + \mu_n) \\
|
||
* &= x_n^2 - \mu_{n-1}^2 + (\mu_{n-1} - x_n)(\mu_{n-1} + \mu_n) \\
|
||
* &= x_n^2 - \mu_{n-1}^2 + \mu_{n-1}^2 - x_n\mu_n - x_n\mu_{n-1} + \mu_n\mu_{n-1} \\
|
||
* &= x_n^2 - x_n\mu_n - x_n\mu_{n-1} + \mu_n\mu_{n-1} \\
|
||
* &= (x_n - \mu_{n-1})(x_n - \mu_n) \\
|
||
* &= S_{n-1} + (x_n - \mu_{n-1})(x_n - \mu_n)
|
||
* \end{align*}
|
||
* ```
|
||
*
|
||
* where we use the identity
|
||
*
|
||
* ```tex
|
||
* x_n - \mu_{n-1} = n (\mu_n - \mu_{n-1})
|
||
* ```
|
||
*
|
||
* ## References
|
||
*
|
||
* - Welford, B. P. 1962. "Note on a Method for Calculating Corrected Sums of Squares and Products." _Technometrics_ 4 (3). Taylor & Francis: 419–20. doi:[10.1080/00401706.1962.10490022](https://doi.org/10.1080/00401706.1962.10490022).
|
||
* - van Reeken, A. J. 1968. "Letters to the Editor: Dealing with Neely's Algorithms." _Communications of the ACM_ 11 (3): 149–50. doi:[10.1145/362929.362961](https://doi.org/10.1145/362929.362961).
|
||
*
|
||
* @param {PositiveInteger} N - number of indexed elements
|
||
* @param {number} correction - degrees of freedom adjustment
|
||
* @param {Float32Array} x - input array
|
||
* @param {integer} stride - stride length
|
||
* @returns {number} variance
|
||
*
|
||
* @example
|
||
* var Float32Array = require( '@stdlib/array/float32' );
|
||
*
|
||
* var x = new Float32Array( [ 1.0, -2.0, 2.0 ] );
|
||
* var N = x.length;
|
||
*
|
||
* var v = svariancewd( N, 1, x, 1 );
|
||
* // returns ~4.3333
|
||
*/
|
||
function svariancewd( N, correction, x, stride ) {
|
||
var delta;
|
||
var mu;
|
||
var M2;
|
||
var ix;
|
||
var v;
|
||
var n;
|
||
var i;
|
||
|
||
n = N - correction;
|
||
if ( N <= 0 || n <= 0.0 ) {
|
||
return NaN;
|
||
}
|
||
if ( N === 1 || stride === 0 ) {
|
||
return 0.0;
|
||
}
|
||
if ( stride < 0 ) {
|
||
ix = (1-N) * stride;
|
||
} else {
|
||
ix = 0;
|
||
}
|
||
M2 = 0.0;
|
||
mu = 0.0;
|
||
for ( i = 0; i < N; i++ ) {
|
||
v = x[ ix ];
|
||
delta = float64ToFloat32( v - mu );
|
||
mu = float64ToFloat32( mu + float64ToFloat32( delta / (i+1) ) );
|
||
M2 = float64ToFloat32( M2 + float64ToFloat32( delta * float64ToFloat32( v - mu ) ) ); // eslint-disable-line max-len
|
||
ix += stride;
|
||
}
|
||
return float64ToFloat32( M2 / n );
|
||
}
|
||
|
||
|
||
// EXPORTS //
|
||
|
||
module.exports = svariancewd;
|