基于聚类粒子群算法网络异常检测模型研究

Model of Intrusion Detection Based on Means Clustering and PSO Algorithm

提出了一种新的基于聚类算法和遗传算法相结合的入侵检测方法模型.算法对聚类的中心采用二进制编码,将网络的正常行为和非正常行为分为不同的类,把每个点到它们之间的各自的聚类中心的欧几里得距离的综合作为相似度量,然后采用粒子群优化算法,有效的降低网络拓扑路径长度,通过优化算法来寻找聚类的中心.Matlab仿真实验结果表明,提出的改进的网络异常检测方法,与较传统网络入侵检测系统模型相比,具有更好的入侵识别率和检测率,同时提高了算法的执行效率.

A new clustering algorithm and genetic algorithm based on a combination of intrusion detection model.Algorithm to cluster the center of the binary code of each point to their respective cluster centers between the Euclidean distance as similarity measure integrated,and then use genetic algorithm,effectively reduce the path length of the network topology,the genetic algorithm to find the cluster center.Matlab simulation results show that the improved network anomaly detection,and the more traditional model of network intrusion detection system,compared with the invasion of a better recognition rate and detection rate,while increasing the efficiency of the algorithm.