Cosine Similarity

Computes the cosine similarity between two arrays.

Cosine similarity defines vector similarity in terms of the angle separating two vectors.

Installation

$ npm install compute-cosine-similarity

For use in the browser, use browserify.

Usage

var similarity = require( 'compute-cosine-similarity' );

similarity( x, y[, accessor] )

Computes the cosine similarity between two arrays.

var x = [ 5, 23, 2, 5, 9 ],
    y = [ 3, 21, 2, 5, 14 ];

var s = similarity( x, y );
// returns ~0.975

For object arrays, provide an accessor function for accessing numeric values.

var x = [
    {'x':2},
    {'x':4},
    {'x':5}
];

var y = [
    [1,3],
    [2,1],
    [3,5]
];

function getValue( d, i, j ) {
    if ( j === 0 ) {
        return d.x;
    }
    return d[ 1 ];
}

var s = similarity( x, y, getValue );
// returns ~0.882

The accessor function is provided three arguments:

d: current datum.
i: current datum index.
j: array index; e.g., array x has index 0, and array y has index 1.

If provided empty arrays, the function returns null.

Examples

var similarity = require( 'compute-cosine-similarity' );

var x = new Array( 100 ),
    y = new Array( 100 ),
    s;

for ( var i = 0; i < x.length; i++ ) {
    x[ i ] = Math.round( Math.random()*100 );
    y[ i ] = Math.round( Math.random()*100 );
}
s = similarity( x, y );

console.log( s );

To run the example code from the top-level application directory,

$ node ./examples/index.js

Tests

Unit

Unit tests use the Mocha test framework with Chai assertions. To run the tests, execute the following command in the top-level application directory:

$ make test

All new feature development should have corresponding unit tests to validate correct functionality.

Test Coverage

This repository uses Istanbul as its code coverage tool. To generate a test coverage report, execute the following command in the top-level application directory:

$ make test-cov

Istanbul creates a ./reports/coverage directory. To access an HTML version of the report,

$ make view-cov

License

MIT license.

Detalhes do pacote

compute-cosine-similarity

readme (leia-me)