Abstract

"Feature selection is a problem of discovering efficient features among all the features in which the final feature set can improve the accuracy and reduce complexity. This is very essential due to rapid escalation in amount of data and information for every 20 months. In some cases, too many redundant or irrelevant features may defeat main features for classification. Feature selection can remedy this problem and therefore improve the prediction accuracy and reduce the computational overhead of classification algorithms. This feature selection methods can be used in many fields like pattern recognition, machine learning, signal processing. Irrelevant features do not contribute to the predictive accuracy, and redundant features do not contribute to getting a better predictor for that they provide most information which is already present in other feature(s). Many feature selection methods have been proposed and studied for machine learning applications. The proposed Bio Inspired Hybrid Bat algorithm for feature selection with Naïve Bayes Classifier (BANB) will select minimum number of relevant features in order to maintain the classification accuracy. This feature selection method is compared against other two algorithms such as Exhaustive Search and Genetic Search. This work focuses on three perspectives: Number of features, classification accuracy and generalization. Results showsthat BANB outperforms against other two feature selection algorithms in selecting lower number of features by removing irrelevant, redundant, or noisy features to maintain highest classification accuracy"

Authors and Affiliations