distributions-poisson-pmf

0.0.2 • Public • Published

Probability Mass Function

NPM version Build Status Coverage Status Dependencies

Poisson distribution probability mass function (PMF).

The probability mass function (PMF) for a Poisson random variable is

Probability mass function (PMF) for a Poisson distribution.

where lambda > 0 is the mean parameter.

Installation

$ npm install distributions-poisson-pmf

For use in the browser, use browserify.

Usage

var pmf = require( 'distributions-poisson-pmf' );

pmf( x[, options] )

Evaluates the probability mass function (PMF) for the Poisson distribution. x may be either a number, an array, a typed array, or a matrix.

var matrix = require( 'dstructs-matrix' ),
    mat,
    out,
    x,
    i;
 
out = pmf( 1 );
// returns ~0.368
 
out = pmf( -1 );
// returns 0
 
out = pmf( 0.5 );
// returns 0
 
= [ 0, 1, 2, 3, 4, 5 ];
out = pmf( x );
// returns [ ~0.368, ~0.368, ~0.184, ~0.0613, ~0.0153, ~0.00307 ]
 
= new Int8Array( x );
out = pmf( x );
// returns Float64Array( [~0.368,~0.368,~0.184,~0.0613,~0.0153,~0.00307] )
 
= new Int16Array( 6 );
for ( i = 0; i < 6; i++ ) {
    x[ i ] = i;
}
mat = matrix( x, [3,2], 'int16' );
/*
    [ 0  1
      2  3
      4  5 ]
*/
 
out = pmf( mat );
/*
    [ ~0.368 ~0.368
      ~0.184 ~0.0613
      ~0.0153 ~0.00307 ]
*/

The function accepts the following options:

  • lambda: mean parameter. Default: 1.
  • accessor: accessor function for accessing array values.
  • dtype: output typed array or matrix data type. Default: float64.
  • copy: boolean indicating if the function should return a new data structure. Default: true.
  • path: deepget/deepset key path.
  • sep: deepget/deepset key path separator. Default: '.'.

A Poisson distribution is a function of one parameter: lambda > 0(mean parameter). By default, lambda is equal to 1. To adjust it, set the corresponding option.

var x = [ 0, 1, 2, 3, 4, 5 ];
 
var out = pmf( x, {
    'lambda': 5,
});
// returns [ ~0.00674, ~0.0337, ~0.0842, ~0.14, ~0.175, ~0.175 ]

For non-numeric arrays, provide an accessor function for accessing array values.

var data = [
    [0,0],
    [1,1],
    [2,2],
    [3,3],
    [4,4],
    [5,5]
];
 
function getValue( d, i ) {
    return d[ 1 ];
}
 
var out = pmf( data, {
    'accessor': getValue
});
// returns [ ~0.368, ~0.368, ~0.184, ~0.0613, ~0.0153, ~0.00307 ]

To deepset an object array, provide a key path and, optionally, a key path separator.

var data = [
    {'x':[0,0]},
    {'x':[1,1]},
    {'x':[2,2]},
    {'x':[3,3]},
    {'x':[4,4]},
    {'x':[5,5]}
];
 
var out = pmf( data, {
    'path': 'x/1',
    'sep': '/'
});
/*
    [
        {'x':[0,~0.368]},
        {'x':[1,~0.368]},
        {'x':[2,~0.184]},
        {'x':[3,~0.0613]},
        {'x':[4,~0.0153]},
        {'x':[5,~0.00307]}
    ]
*/
 
var bool = ( data === out );
// returns true

By default, when provided a typed array or matrix, the output data structure is float64 in order to preserve precision. To specify a different data type, set the dtype option (see matrix for a list of acceptable data types).

var x, out;
 
= new Int8Array( [0,1,2,3,4] );
 
out = pmf( x, {
    'dtype': 'float32'
});
// returns Float32Array( [~0.00674,~0.0337,~0.0842,~0.14,~0.175] )
 
// Works for plain arrays, as well...
out = pmf( [0,1,2,3,4], {
    'dtype': 'float32'
});
// returns Float32Array( [~0.00674,~0.0337,~0.0842,~0.14,~0.175] )

By default, the function returns a new data structure. To mutate the input data structure (e.g., when input values can be discarded or when optimizing memory usage), set the copy option to false.

var bool,
    mat,
    out,
    x,
    i;
 
= [ 0, 1, 2, 3, 4, 5 ];
 
out = pmf( x, {
    'copy': false
});
// returns [ ~0.368, ~0.368, ~0.184, ~0.0613, ~0.0153, ~0.00307 ]
 
bool = ( x === out );
// returns true
 
= new Int16Array( 6 );
for ( i = 0; i < 6; i++ ) {
    x[ i ] = i;
}
mat = matrix( x, [3,2], 'int16' );
/*
    [ 0  1
      2  3
      4  5 ]
*/
 
out = pmf( mat, {
    'copy': false
});
/*
    [ ~0.368 ~0.368
      ~0.184 ~0.0613
      ~0.0153 ~0.00307 ]
*/
 
bool = ( mat === out );
// returns true

Notes

  • If an element is not a numeric value, the evaluated PMF is NaN.

    var data, out;
     
    out = pmf( null );
    // returns NaN
     
    out = pmf( true );
    // returns NaN
     
    out = pmf( {'a':'b'} );
    // returns NaN
     
    out = pmf( [ true, null, [] ] );
    // returns [ NaN, NaN, NaN ]
     
    function getValue( d, i ) {
        return d.x;
    }
    data = [
        {'x':true},
        {'x':[]},
        {'x':{}},
        {'x':null}
    ];
     
    out = pmf( data, {
        'accessor': getValue
    });
    // returns [ NaN, NaN, NaN, NaN ]
     
    out = pmf( data, {
        'path': 'x'
    });
    /*
        [
            {'x':NaN},
            {'x':NaN},
            {'x':NaN,
            {'x':NaN}
        ]
    */
  • Be careful when providing a data structure which contains non-numeric elements and specifying an integer output data type, as NaN values are cast to 0.

    var out = pmf( [ true, null, [] ], {
        'dtype': 'int8'
    });
    // returns Int8Array( [0,0,0] );

Examples

var pmf = require( 'distributions-poisson-pmf' ),
    matrix = require( 'dstructs-matrix' );
 
var data,
    mat,
    out,
    tmp,
    i;
 
// Plain arrays...
data = new Array( 10 );
for ( i = 0; i < data.length; i++ ) {
    data[ i ] = i * 0.5;
}
out = pmf( data );
 
// Object arrays (accessors)...
function getValue( d ) {
    return d.x;
}
for ( i = 0; i < data.length; i++ ) {
    data[ i ] = {
        'x': data[ i ]
    };
}
out = pmf( data, {
    'accessor': getValue
});
 
// Deep set arrays...
for ( i = 0; i < data.length; i++ ) {
    data[ i ] = {
        'x': [ i, data[ i ].x ]
    };
}
out = pmf( data, {
    'path': 'x/1',
    'sep': '/'
});
 
// Typed arrays...
data = new Int32Array( 10 );
for ( i = 0; i < data.length; i++ ) {
    data[ i ] = i;
}
out = pmf( data );
 
// Matrices...
mat = matrix( data, [5,2], 'int32' );
out = pmf( mat );
 
// Matrices (custom output data type)...
out = pmf( mat, {
    'dtype': 'uint8'
});

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.

Copyright

Copyright © 2015. The Compute.io Authors.

Package Sidebar

Install

npm i distributions-poisson-pmf

Weekly Downloads

2

Version

0.0.2

License

MIT

Last publish

Collaborators

  • kgryte
  • planeshifter