A Multi-Population Univariate Marginal Distribution Algorithm for Feature Selection

Feature selection is to select an informative subset from original feature set aiming at reducing the dimension of the feature space and enhancing the performance of the classifier. It is a crucial problem of pattern recognition and has attracted much attention in recent years. In this paper, we have proposed a multi-population univariate marginal distribution algorithm using random population to increase the diversity to avoid falling into local optima. The experimental results showed that the proposed algorithm could effectively improve the accuracy of the classifier, at the same time reduce the dimension of the features.