Formulation of Feature Selection with Support Vector Machine

Abstract

Basic question arises when classification came in picture classification accuracy, ensemble size, and computational complexity. Feature selection is importance for improvement and performance of classification algorithm. Classification algorithm may not scale up to the size of the full feature set either in sample or time but with feature selection help us to better understand the domain with Cheaper to collect a subset of predictors and Safer to collect a reduced subset of predictors. An important pre-processing step in classification tasks is called as, Feature selection its aims to minimize both the classification error rate and the number of features for inference knowledge in any domain. Feature selection is Minimum set F that achieves maximum classification performance of T (for a given set of classifiers and classification performance metrics). This paper proposes feature selection methodology which includes ranking, information gain and filter method concept. After the feature subset train SVM with RBF kernel for classification.