culebra.fitness_function module

Fitness functions.

This module provides several fitness functions, all related to the feature selection problem. Currently:

  • The abc sub-module provides abstract classes to define the remaining fitness functions.

  • The dataset_score sub-module provides abstract classes to score dataset-related problems.

  • The feature_selection sub-module is centered in datasets dimensionality reduction.

  • The svc_optimization sub-module provides several fitness functions intended to optimize the Support Vector Classifier (SVC) hyperparameters for a given dataset.

  • The cooperative sub-module provides fitness functions designed to the cooperative solving of a feature selection problem while the classifier hyperparamters are also being optimized.

  • The tsp sub-module offers fitness functions for the traveling salesman problem.

Attributes

DEFAULT_CLASSIFIER = <class 'sklearn.naive_bayes.GaussianNB'>

Default classifier for fitness functions

DEFAULT_CV_FOLDS = 5

Default number of folds for cross-validation

DEFAULT_THRESHOLD = 0

Default similarity threshold for fitnesses