Unsupervised Classification Algorithm for Intrusion Detection based on Competitive Learning Network

Classification of intrusion attacks and normal network traffic is a challenging and critical problem in network security. Many classification methods for intrusion detection have been proposed, but there are few algorithms that are capable of distinguishing among the various attacks and normal connections effectively. This paper presents an effective intrusion detection algorithm based on conscientious rival penalized competitive learning (CRPCL), which improves RPCL to set a conscientious threshold to restrict a winner that won too many times and to make every neural unit win the competition at near ideal probability. To assess the classification performance of the algorithm, it is compared with some well-known classifiers. The experiments with KDD CUP 99 data indicate that this method has good performance and can improve the detection quality effectively.