Multi-Strategies Ensemble Based On Artificial Immune System For Intrusion Detection System

In recent years, with the widely application of the computer network in our daily life, Network security issues are becoming increasingly attracted human’s attention and concern. In the safe system construction of network, the Firewall is the center and base of all the security linkage technology, but the Firewall cannot meet the demand for network security anymore, so Intrusion detection system has become an effective complement to Firewalls. Intrusion detection is a technology which is designed and configured to detect and report unauthorized or unusual phenomenon to ensure the security of the computer systems. At the same time, the body’s immune system has many features, such as identifying and removing the external antigens and so on. The antigen recognition and the removal process in the immune system are very similar to the process of the Intrusion detection. We take the immune mechanism into the intrusion detection technology of network security, and it will be able to detect anomalies and abnormal data in network traffic. Based on the research of the immune danger theory, the dendritic cells in immunology has been studied and improved and finally we combine with the negative selection algorithm, non-negative matrix algorithm and the dendritic cell algorithm and propose Multi-Strategies Ensemble based on Artificial Immune Systems for Intrusion Detection.First of the dissertation, we introduce the deterministic parameters and dynamic optimization mechanism into the dendritic cells algorithm and apply the improved dendritic cells algorithm to the intrusion detection systems. Compared to the negative selection algorithm that needs many training samples and the traditional dendritic cells algorithm with multi-parameters, the improved algorithm can effectively improve the detection rate, reduce false alarm rate and steady the results of test.In order to make the detection rate better, we propose the strategy of best of three sets and design a mode with Multi-Strategies Ensemble based on Artificial Immune Systems for Intrusion Detection. We combine with the negative selection algorithm, non-negative matrix algorithm and the dendritic cell algorithm. According to the simulation results, the system can effectively improve the detection rate, reduce false alarm rate and steady the results, which is a good way to intrusion detection.