A Feature Selection Method Base on Fisher Score for Detecting SVM-Based Network Intrusions

The diagnostic process is based on the fact that malicious activity is different from the activity of a normal system. Detection of intrusion is a very complex process. In this paper, we propose Feature Selection to improve the velocity support vector machines (SVM) based intrusion detection system (IDS). The new model has used a feature selection method based on Fisher Score with an innovation in fitness function reduce the dimension of the data, increase true positive detection and simultaneously decrease false positive detection. In addition, the computation time for training will also have a remarkable reduction. We demonstrate the feasibility of our method by performing several experiments on NSLKDD intrusion detection system competition dataset. Results show that the proposed method can reach high accuracy and low false positive rate (FPR) simultaneously. Numeric Results and comparison to other models have been presented.