Feature Subset Selection Using Genetic Algorithm with Aggressive Mutation for Classification Problem

Abstract

In this work, the algorithmic approaches to Feature Subset Selection (FSS) are reviewed. FSS is the technique of selecting a subset of relevant features for building parsimonious models. One successful algorithm in selecting relevant features in Brain-Computing Interface is the Genetic Algorithm with Aggressive Mutation (GAAM). We implemented a scikit-learn compatible library for GAAM and determined its applicability with classification tasks in general. Identifying relevant features in a predictive modeling task improves the interpretability of the model, reduces its complexity and the time requirement for training.