Classifier Model for Intrusion Detection Using Bio-inspired Metaheuristic Approach

In machine learning and statistics, feature selection is the technique of selecting a subset of relevant features for building robust learning models. In this paper we propose a bio-inspired BAT algorithm as feature selection method to find the optimal features from the KDDCup’99 intrusion detection dataset obtained from UCI Machine Learning repository. Neural Networks (NN) as a classifier collects data randomly from the dataset and constructs a training dataset with original records using bagging approach. The performance of Neural Network with Repeated Incremental Pruning to Produce Error Reduction (RIPPER) is compared with the Neural Network with C4.5 decision tree and the experimental result shows that the Neural Network with RIPPER outperforms the other algorithm. Keywords— Bio-inspired, RIPPER, bagging, decision tree, KDDCup’99..