A Proactive Model for Optimizing Swarm Search Algorithms for Intrusion Detection System

Abstract

This paper proposes a novel a Proactive model for Swarm Optimization for feature selection (PMSO) of intrusion detection system. The proposed model potentially enhances the search process and initial value of population for swarm optimization algorithms. The swarm optimization often suffering from falling in a local optimum, one of the main causes of this problem is the initial value of the population. Our proposed model is applied with the aim to address these challenges by restarting generate the population when system sensed the search process go forward stagnation. This process will give high exploration and reduce the probability of Stagnation in local optima. Benchmark datasets, namely, NSL-KDD datasets is used to demonstrate and validate the performance of the proposed model for intrusion detection. Experimentally, the machine learning accuracy has provided better classification results with proposed system (%2) than swarm optimization (Particle Swarm Optimization and Bat Algorithm).