2008 | OriginalPaper | Buchkapitel
Genetic Programming for Feature Ranking in Classification Problems
verfasst von : Kourosh Neshatian, Mengjie Zhang, Peter Andreae
Erschienen in: Simulated Evolution and Learning
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
Feature ranking (FR) provides a measure of usefulness for the attributes of a classification task. Most existing FR methods focus on the relevance of a single feature to the class labels. Here, we use GP to see how a set of features can contribute towards discriminating different classes and then we score the participating features accordingly. The scoring mechanism is based on the frequency of appearance of each feature in a collection of GP programs and the fitness of those programs. Our results show that the proposed FR method can detect important features of a problem. A variety of different classifiers restricted to just a few of these high-ranked features work well. The ranking mechanism can also shrink the search space of size
O
(2
n
) of subsets of features to a search space of size
O
(
n
) in which there are points that may improve the classification performance.