Feature Selection based F-score and ACO Algorithm in Support Vector Machine

This study proposes a new strategy combining with the SVM(support vector machine) classifier for features selection that retains sufficient information for classification purpose. Our proposed approach uses F-score models to optimize feature space by removing both irrelevant and redundant features. To improve classification accuracy, the parameters optimization of the penalty constant C and the bandwidth of the radial basis function (RBF) kernel gamma is an important step in establishing an efficient and high-performance support vector machine (SVM) model. Aiming at optimizing the parameters of SVM, this paper also presents a grid based ant colony optimization (ACO) algorithm to choose parameters C and gamma automatically for SVM instead of selecting parameters randomly by human's experience and traditional grid searching algorithm, so that the classification feature numbers can be reduced and the classification performance can be improved simultaneously. Some experimental results confirm the feasibility and efficiency of the approach.