REVISTA DYNA NEW TECHNOLOGIES

Key words:

Operador de alteración, algoritmo genético, algoritmo de estimación de la distribución, teoría de testores, selección de subconjuntos de características

Article type:

ARTICULO DE INVESTIGACION / RESEARCH ARTICLE

Section:

RESEARCH ARTICLES

ABSTRACT:
This paper presents the design and integration of the “alteration operator” to evolutive metaheuristics (GA and EDA). The aim is searching typical testors for feature selection in medical data. This operator avoids the aleatory search working as a support of the exploration of the solutions space. The hybrid metaheuristics found the typical testors associated to the analyzed data which are irreducible subsets that allow to classify new instances. The informational weight was used to know the impact of each feature in the classification process. Finally, the tuning and contrasting of metaheuristics is presented in order to find the best parameter values that improve their performance and determine the most appropriate metaheuristics for each pathology.

Keywords: alteration operator, genetic algorithm, estimation of distribution algorithm, testor theory, feature subset selection.