Artificial Neural Network Classifier for Intrusion Detection System in Computer Network

An intrusion detection system is a security management tool for computers and networks. An intrusion is mainly a try to violation of network security requirements and norms. Detection deals with the countermeasures to detect such attacks. The goal of the intrusion detection system mechanism is to observe the network traffic if any packet whose pattern varies when standard to the normal behavior is said to be an anomaly and hence an attack. The main objective of this paper is to perform data preprocessing on KDD CUP 99 dataset to select a subset of features to advance the speed of the detection process. A modified Kolmogorov-Smirnov correlation-based filter algorithm is used to select features. And propose an intrusion detection model using PSO-WENN; this can classify the attacks effectively and reduce the number of false alarms generated by an intrusion detection system and improve the attack detection rate.