time-to-botec/squiggle/node_modules/@stdlib/stats/ranks/README.md
NunoSempere b6addc7f05 feat: add the node modules
Necessary in order to clearly see the squiggle hotwiring.
2022-12-03 12:44:49 +00:00

4.8 KiB

ranks

Compute ranks for values of an array-like object.

Usage

var ranks = require( '@stdlib/stats/ranks' );

ranks( arr[, opts] )

Returns the sample ranks of the elements in arr, which can be either an array or typed array.

var arr = [ 1.1, 2.0, 3.5, 0.0, 2.4 ];
var out = ranks( arr );
// returns [ 2, 3, 5, 1, 4 ]

// Ties are averaged:
arr = [ 2, 2, 1, 4, 3 ];
out = ranks( arr );
// returns [ 2.5, 2.5, 1, 5, 4 ];

// Missing values are placed last:
arr = [ null, 2, 2, 1, 4, 3, NaN, NaN ];
out = ranks( arr );
// returns [ 6, 2.5, 2.5, 1, 5, 4, 7 ,8 ]

The function accepts the following options:

  • method: string indicating how ties are handled. Can be one of the following values: 'average', 'min', 'max', 'ordinal' and 'dense'. Default: 'average'.
  • missing: string specifying how missing values are handled. Must be either 'last', 'first' or 'remove'. Default: 'last'.
  • encoding: array holding all values which will be regarded as missing values. Default: [ NaN, null].

When all elements of the array are different, the ranks are uniquely determined. When there are equal elements (called ties), the method option determines how they are handled. The default, 'average', replace the ranks of the ties by their mean. Other possible options are 'min' and 'max', which replace the ranks of the ties by their minimum and maximum, respectively. 'dense' works like 'min', with the difference that the next highest element after a tie is assigned the next smallest integer. Finally, ordinal gives each element in arr a distinct rank, according to the position they appear in.

var data = [ 2, 2, 1, 4, 3 ];

// Max method:
var out = ranks( data, {
    'method': 'max'
});
// returns [ 3, 3, 1, 5, 4 ]

// Min method:
out = ranks( data, {
    'method': 'min'
});
// returns [ 2, 2, 1, 5, 4 ]

// Ordinal method
out = ranks( data, {
    'method': 'ordinal'
});
// returns [ 2, 3, 1, 5, 4 ]

// Dense method:
out = [ 2, 2, 1, 4, 3 ];
out = ranks( data, {
    'method': 'dense'
});
// returns [ 2, 2, 1, 4, 3 ]

The missing option is used to specify how to handle missing data. By default, NaN or null are treated as missing values. 'last'specifies that missing values are placed last, 'first' that the are assigned the lowest ranks and 'remove' means that they are removed from the array before the ranks are calculated.

var data = [ NaN, 2, 2, 1, 4, 3, null, null ];

var out = ranks( data, {
    'missing': 'first'
});
// returns [ 1, 5.5, 5.5, 4, 8, 7, 2, 3 ]

out = ranks( data, {
    'missing': 'last'
});
// returns [ 6, 2.5, 2.5, 1, 5, 4, 7 ,8 ]

out = ranks( data, {
    'missing': 'remove'
});
// returns [ 2.5, 2.5, 1, 5, 4 ]

Custom encoding for missing values is supported via the encoding option, which allows to supply the function with an array of values which should be treated as missing.

var Int32Array = require( '@stdlib/array/int32' );

var data = new Int32Array( [ 2, 1, -999, 3, 4 ] );

var out = ranks( data, {
    'encoding': [ -999 ]
});
// returns [ 2, 1, 5, 3, 4 ]

Examples

var Int32Array = require( '@stdlib/array/int32' );
var round = require( '@stdlib/math/base/special/round' );
var randu = require( '@stdlib/random/base/randu' );
var ranks = require( '@stdlib/stats/ranks' );

var data;
var out;
var i;

// Plain arrays...
data = new Array( 10 );
for ( i = 0; i < data.length; i++ ) {
    data[ i ] = round( randu()*10.0 );
}

out = ranks( data );
// returns <array>

// Typed arrays...
data = new Int32Array( 10 );
for ( i = 0; i < data.length; i++ ) {
    data[ i ] = randu() * 10.0;
}

out = ranks( data );
// returns <array>