skcriteria.validate module

This module core functionalities for validate the data used inside scikit criteria.

  • Constants that represent minimization and mazimization criteria.
  • Scikit-Criteria Criteria ndarray creation.
  • Scikit-Criteria Data validation.
skcriteria.validate.MIN = -1

Int: Minimization criteria

skcriteria.validate.MAX = 1

Int: Maximization criteria

exception skcriteria.validate.DataValidationError[source]

Bases: exceptions.ValueError

Raised when some part of the multicriteria data (alternative matrix, criteria array or weights array) are not compatible with another part.

skcriteria.validate.criteriarr(criteria)[source]

Validate if the iterable only contains MIN (or any alias) and MAX (or any alias) values. And also always returns an ndarray representation of the iterable.

Parameters:

criteria : Array-like

Iterable containing all the values to be validated by the function.

Returns:

numpy.ndarray :

Criteria array.

Raises:

DataValidationError :

if some value of the criteria array are not MIN (-1) or MAX (1)

skcriteria.validate.validate_data(mtx, criteria, weights=None)[source]

Validate if the main components of the Data in scikit-criteria are compatible.

The function tests:

  • The matrix (mtx) must be 2-dimensional.
  • The criteria array must be a criteria array (criteriarr function).
  • The number of criteria must be the same number of columns in mtx.
  • The weight array must be None or an iterable with the same length of the criteria.
Parameters:

mtx : 2D array-like

2D alternative matrix, where every column (axis 0) are a criteria, and every row (axis 1) is an alternative.

criteria : Array-like

The sense of optimality of every criteria. Must has only MIN (-1) and MAX (1) values. Must has the same elements as columns has mtx

weights : array like or None

The importance of every criteria. Must has the same elements as columns has mtx or None.

Returns:

mtx : numpy.ndarray

mtx representations as 2d numpy.ndarray.

criteria : numpy.ndarray

A criteria as numpy.ndarray.

weights : numpy.ndarray or None

A weights as numpy.ndarray or None (if weights is None).

Raises:

DataValidationError :

If the data are incompatible.