wowa 1.12

Calculate weighted OWA functions and extending bivariate means

wowa

This package calculates weighted OWA functions and extending bivariate means" Functions are:

• py_WAM: WOWATree callback function if sorting is needed in general
• py_OWA: WOWATree callback function if no sorting is needed when used in the tree
• WOWATree: symmetric base aggregator
• WAn: processes the tree
• weightedOWAQuantifierBuild: calculates spline knots and coefficients for later use in weightedOWAQuantifier
• weightedOWAQuantifier: Calculates the value of the WOWA, with quantifier function obtained in weightedOWAQuantifierBuild
• ImplicitWOWA: Calculates implicit Weighted OWA function
• WAM: weighted arithmetic mean function
• OWA: ordered weighted averaging function

Documentation

User Manual

Installation

To install type:

$ pip install wowa

Usage of py_OWA( n, x, w)

from wowa import py_OWA

WOWATree callback function if sorting is needed in general

Parameters

Input parameters:

n: size of arrays
x[]: NumPy array of size n, float
w[]: NumPy array of size n, float

Output parameters:

double y: aggregated sum

Usage of py_WAM( n, x, w)

from wowa import py_WAM

WOWATree callback function if no sorting is needed when used in the tree

Parameters

Input parameters:

n: size of arrays
x[]: NumPy array of size n, float
w[]: NumPy array of size n, float

Output parameters:

double y: aggregated sum

Usage of WOWATree( x, p, w, cb, L)

from wowa import WOWATree

Symmetric base aggregator. The weights must add to one and be non-negative.

Parameters

Input parameters:

x[]: NumPy array of inputs, size n, float
p[]: NumPy array of weights of inputs x[], size n, float
w[]: NumPy array of weights for OWA, size n, float
cb: Nallback function. Either pre-defined py_OWA() or py_WAM() or user defined of type float(ch*)(int, float[], float[], int)
L: Number of binary tree levels. Run time = O[(n-1)L]

Output parameters:

y: weightedf, double

Usage of WAn( x, w, L, F)

from wowa import WAn

Parameters

Input parameters:

x[]: NumPy array of inputs, size n, float
w[]: NumPy array of weights for OWA, size n, float
L: Number of binary tree levels
F: User defined callback function of type float(*F)( float, float)

Output parameters:

y: result of tree processing, double

Usage of weightedOWAQuantifierBuild( p, w)

from wowa import weightedOWAQuantifierBuild

Parameters

Input parameters:

p[]: NumPy array of weights of inputs x[], size n, float
w[]: NumPy array of weights for OWA, size n, float

Output parameters:

spline: the spline knots and coefficients for later use in weightedOWAQuantifier
T: the number of knots in the monotone spline

Usage of weightedOWAQuantifier( x, p, w, spline, T);

from wowa import weightedOWAQuantifier

Calculates the value of the WOWA, with quantifier function obtained in weightedOWAQuantifierBuild

Parameters

Input parameters:

x[]: NumPy array of inputs, size n, float
p[]: NumPy array of weights of inputs x[], size n, float
w[]: NumPy array of weights for OWA, size n, float
spline[]: keeps the spline knots and coefficients. Output from weightedOWAQuantifierBuild
T: the number of knots in the monotone spline

Output parameters:

y: double

Usage of ImplicitWOWA x, p, w)

from wowa import ImplicitWOWA

Calculates implicit Weighted OWA function

Parameters

Input parameters:

x[]: NumPy array of inputs, size n, float
p[]: NumPy array of weights of inputs x[], size n, float
w[]: NumPy array of weights for OWA, size n, float

Output parameters:

y: float

Usage of WAM( n, x, w)

from wowa import WAM

Weighted arithmetic mean function

Parameters

Input parameters:

n: size of arrays
x[]: NumPy array of size n, float
w[]: NumPy array of size n, float

Output parameters:

y: aggregated sum, float

Usage of OWA( n, x, w)

from wowa import OWA

Ordered weighted averaging function

Parameters

Input parameters:

n: size of arrays
x[]: NumPy array of size n, float
w[]: NumPy array of size n, float

Output parameters:

y: aggregated sum, float

Test

To unit test type:

$ test/test.py