Genetic algorithms in feature selection

We use a genetic algorithm (GA) for the feature selection problem. The method explores the space of possible subsets to obtain the set of features that maximizes the predictive accuracy and minimizes irrelevant attributes. We introduce a multiple correlation in a fitness function used by the GA to evaluate the fitness of each feature subset regarding relationship in its domain. Comparison between our fitness function and the traditional fitness function is done on five problem domains. The empirical results demonstrate that the proposed fitness function is more effective compared to the traditional fitness function on all cases considered.