Feature Selection Using Memetic Algorithms

The feature selection process can be considered a problem of global combinatorial optimization in machine learning, which reduces the number of features, removes irrelevant, noisy and redundant data, and results in acceptable classification accuracy. In this study, we propose a combined filter method (ReliefF) and a wrapper method (memetic algorithm, MA) for classification. The goal of our method is to filter the irrelevant features and select the most important feature subsets. We used the ReliefF algorithm to calculate and update the scores of every feature for each data set, and then applied a MA for feature selection. The K-nearest neighbor (K-NN) method with leave-one-out cross-validation (LOOCV) serves as a classifier for evaluating classification accuracies. The experimental results show that the proposed method is superior to existing methods in terms of classification accuracy.