Network Intrusion Detection System Based on Optimized Fuzzy Rules Algorithm

As computer networks and distributed applications more complex, diverse and intelligent, network behavior anomaly detection has gradually become the effective monitoring and system controlling technology. The paper established a network intrusion detection system and to investigate the data rules, sensors and abnormal behavior automatic identification in this system, a kind of algorithm based on fuzzy rules to describe the network abnormal behavior was introduced into this paper. It was used to describe the misclassification invasion rules effectively and then to convert the misclassification invasion rules to the issue of seeking optimal separating hyper plane. Subsequently, the double super ball membership function was introduced into the system to restrict the intrusion features, and to establish intrusion rule set which was used to make optimized description of the intrusion rule set and then complete intrusion detection. The experimental results showed that: in the context of different network attacks, the system can complete a variety of attacks and efficient detection. The detection error was not more than 1% which basically met the requirements of the reliable, high precision, anti-interference ability in automatic network intrusion detection and provided a reference to the future research on network intrusion detection.