time-to-botec/squiggle/node_modules/@stdlib/stats/wilcoxon/lib/main.js

312 lines
8.3 KiB
JavaScript
Raw Normal View History

/* eslint-disable max-statements, max-lines-per-function */
/**
* @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 isNumberArray = require( '@stdlib/assert/is-number-array' ).primitives;
var isTypedArrayLike = require( '@stdlib/assert/is-typed-array-like' );
var setReadOnly = require( '@stdlib/utils/define-read-only-property' );
var isObject = require( '@stdlib/assert/is-plain-object' );
var ranks = require( './../../ranks' );
var normalCDF = require( './../../base/dists/normal/cdf' ).factory;
var signrankCDF = require( './../../base/dists/signrank/cdf' );
var tabulate = require( '@stdlib/utils/tabulate' );
var signum = require( '@stdlib/math/base/special/signum' );
var sqrt = require( '@stdlib/math/base/special/sqrt' );
var abs = require( '@stdlib/math/base/special/abs' );
var Float64Array = require( '@stdlib/array/float64' );
var validate = require( './validate.js' );
var unique = require( './unique.js' );
var print = require( './print.js' ); // eslint-disable-line stdlib/no-redeclare
// VARIABLES //
var pnorm = normalCDF( 0.0, 1.0 );
// MAIN //
/**
* Computes a Wilcoxon signed rank test.
*
* @param {NumericArray} x - data array
* @param {NumericArray} [y] - optional paired data array
* @param {Options} [options] - function options
* @param {number} [options.alpha=0.05] - significance level
* @param {string} [options.alternative='two-sided'] - alternative hypothesis (`two-sided`, `less`, or `greater`)
* @param {string} [options.zeroMethod='wilcox'] - method governing how zero-differences are handled (`pratt`, `wilcox`, or `zsplit`)
* @param {boolean} [options.correction=true] - continuity correction adjusting the Wilcoxon rank statistic by 0.5 towards the mean
* @param {boolean} [options.exact=false] - whether to force using the exact distribution instead of a normal approximation when there are more than fifty data points
* @param {number} [options.mu=0] - location parameter under H0
* @throws {TypeError} `x` must be a numeric array
* @throws {TypeError} `y` must be a numeric array
* @throws {TypeError} options has to be simple object
* @throws {TypeError} `alpha` option has to be a number primitive
* @throws {RangeError} `alpha` option has to be a number in the interval `[0,1]`
* @throws {TypeError} `alternative` option has to be a string primitive
* @throws {Error} `alternative` option must be `two-sided`, `less`, or `greater`
* @throws {TypeError} `zeroMethod` option has to be a string primitive
* @throws {Error} `zeroMethod` option must be `pratt`, `wilcox`, or `zsplit`
* @throws {TypeError} `correction` option has to be a boolean primitive
* @throws {TypeError} `exact` option has to be a boolean primitive
* @throws {TypeError} `mu` option has to be a number primitive
* @returns {Object} test result object
*
* @example
* var x = [ 6, 8, 14, 16, 23, 24, 28, 29, 41, -48, 49, 56, 60, -67, 75 ];
* var out = wilcoxon( x, {
* 'mu': 2
* });
*
* @example
* var x = [ 6, 8, 14, 16, 23, 24, 28, 29, 41, -48, 49, 56, 60, -67, 75 ];
* var out = wilcoxon( x, {
* 'alternative': 'greater'
* });
*/
function wilcoxon() {
var correction;
var zeroMethod;
var options;
var hasTies;
var counts;
var repsum;
var rplus;
var nzero;
var rzero;
var alpha;
var pval;
var opts;
var stat;
var alt;
var err;
var len;
var tmp;
var out;
var ad;
var mu;
var mn;
var se;
var d;
var i;
var r;
var T;
var v;
var x;
var y;
x = arguments[ 0 ];
if ( !isTypedArrayLike( x ) && !isNumberArray( x ) ) {
throw new TypeError( 'invalid argument. First argument must be a numeric array. Value: `' + x + '`.' );
}
len = x.length;
if ( arguments.length > 1 ) {
if ( isObject( arguments[ 1 ] ) ) {
options = arguments[ 1 ];
} else {
y = arguments[ 1 ];
if ( !isTypedArrayLike( y ) && !isNumberArray( y ) ) {
throw new TypeError( 'invalid argument. `y` argument must be a numeric array. Value: `' + y + '`.' );
}
if ( len !== y.length ) {
throw new Error( 'invalid arguments. The first and second arguments must have the same length.' );
}
if ( arguments.length > 2 ) {
options = arguments[ 2 ];
}
}
}
opts = {};
if ( options ) {
err = validate( opts, options );
if ( err ) {
throw err;
}
}
mu = opts.mu || 0.0;
if ( opts.correction === void 0 ) {
correction = true;
} else {
correction = opts.correction;
}
if ( opts.alpha === void 0 ) {
alpha = 0.05;
} else {
alpha = opts.alpha;
}
if ( len < 2 ) {
throw new Error( 'invalid argument. First argument must contain at least two elements. Value: `' + x + '`' );
}
alt = opts.alternative || 'two-sided';
zeroMethod = opts.zeroMethod || 'wilcox';
if ( zeroMethod === 'wilcox' ) {
// Only keep all non-zero differences:
d = [];
if ( y ) {
for ( i = 0; i < len; i++ ) {
v = ( x[ i ] - y[ i ] ) - mu;
if ( v !== 0 ) {
d.push( v );
}
}
} else {
for ( i = 0; i < len; i++ ) {
if ( x[ i ] !== 0 ) {
d.push( x[ i ] - mu );
}
}
}
nzero = x.length - d.length;
} else {
d = new Float64Array( len );
nzero = 0;
if ( y ) {
for ( i = 0; i < len; i++ ) {
d[ i ] = ( x[ i ] - y[ i ] ) - mu;
if ( d[ i ] === 0 ) {
nzero += 1;
}
}
} else {
for ( i = 0; i < len; i++ ) {
d[ i ] = x[ i ] - mu;
if ( d[ i ] === 0 ) {
nzero += 1;
}
}
}
}
if ( nzero === len ) {
throw new Error( '`x` or `x - y` cannot be zero for all elements.' );
}
// Update length after potentially discarding zero values:
len = d.length;
ad = new Float64Array( len );
for ( i = 0; i < len; i++ ) {
ad[ i ] = abs( d[ i ] );
}
r = ranks( ad );
rplus = 0;
rzero = 0;
for ( i = 0; i < len; i++ ) {
if ( d[ i ] > 0 ) {
rplus += r[ i ];
}
else if ( d[ i ] === 0 ) {
rzero += r[ i ];
}
}
hasTies = unique( r ).length !== r.length;
if ( zeroMethod === 'zsplit' ) {
rplus += rzero / 2.0;
}
T = rplus;
mn = len * ( len + 1.0 ) * 0.25;
se = len * ( len + 1.0 ) * ( ( 2.0 * len ) + 1.0 );
if ( zeroMethod === 'pratt' ) {
tmp = [];
for ( i = 0; i < len; i++ ) {
if ( d[ i ] !== 0 ) {
tmp.push( r[ i ] );
}
}
r = tmp;
mn -= nzero * ( nzero + 1.0 ) * 0.25;
se -= nzero * ( nzero + 1.0 ) * ( ( 2.0 * nzero ) + 1.0 );
}
counts = tabulate( r );
repsum = 0;
for ( i = 0; i < counts.length; i++ ) {
if ( counts[ i ][ 1 ] > 1 ) {
v = counts[ i ][ 1 ];
repsum += v * ( (v*v) - 1 );
}
}
if ( repsum > 0 ) {
// Correction for repeated values:
se -= 0.5 * repsum;
}
se = sqrt( se / 24.0 );
if (
( len > 50 && !opts.exact ) ||
nzero > 0 ||
hasTies
) {
d = 0.0;
if ( correction ) {
switch ( alt ) {
case 'two-sided':
d = 0.5 * signum( T - mn );
break;
case 'less':
d = -0.5;
break;
default:
d = 0.5;
break;
}
}
// Compute test statistic and p-value using normal approximation:
stat = ( T - mn - d ) / se;
if ( alt === 'two-sided' ) {
pval = 2.0 * ( 1.0 - pnorm( abs( stat ) ) );
} else if ( alt === 'greater' ) {
pval = 1.0 - pnorm( stat );
} else {
pval = pnorm( stat );
}
} else {
// Compute test statistic and p-value using exact critical values:
stat = T;
if ( alt === 'two-sided' ) {
if ( stat > ( len * ( len+1 ) / 4 ) ) {
pval = 2.0 * ( 1 - signrankCDF( stat - 1, len ) );
} else {
pval = 2.0 * signrankCDF( stat, len );
}
} else if ( alt === 'greater' ) {
pval = 1.0 - signrankCDF( stat - 1, len );
} else {
pval = signrankCDF( stat, len );
}
}
out = {};
setReadOnly( out, 'rejected', pval <= alpha );
setReadOnly( out, 'alpha', alpha );
setReadOnly( out, 'pValue', pval );
setReadOnly( out, 'statistic', T );
setReadOnly( out, 'nullValue', mu );
setReadOnly( out, 'alternative', alt );
setReadOnly( out, 'method', ( ( y ) ? 'Paired' : 'One-Sample' ) + ' Wilcoxon signed rank test' );
setReadOnly( out, 'print', print );
return out;
}
// EXPORTS //
module.exports = wilcoxon;