粒子群优化聚类算法在网络安全中的应用

Application of Particle Swarm Optimization Clustering in Network Security

研究快速变异的网络攻击准确检测问题,网络攻击如果在入侵过程中,发生较快的变异,使得入侵特征很难被准确的描述。传统的C均值聚类(FCM)算法在网络入侵检测中,多是依靠特征匹配完成检测,由于无法准确描述快速变异的入侵特征,导致网络入侵初始聚类中心选择不当,检测正确率不高。提出一种粒子群优化聚类算法的网络入侵检测方法,通过粒子群算法选择初始聚类中心,检测变异后入侵的最小化特征,采用FCM算法对最小特征进行聚类分析,完成快速变异网络入侵的检测。仿真结果表明,改进FCM算法能很好克服传统FCM算法的缺陷,有效地提高了网络检测正确率,同时提高了网络入侵的检测速度。

To study the variation of network attack fast accurate detection problem, the paper put forward a particle swarm optimization clustering algorithm of network intrusion detection methods. First, through the particle swarm algorithm, the initial clustering center was selected. After the invasion detection variation characteristics minimization, the minimum FCM algorithm Characteristics of clustering analysis was used to the complete network intrusion detection fast variation. The simulation results show that the improved FCM algorithm is very good in overcoming the defects of the traditional FCM algorithm, it can effectively improve the accuracy of network detection, and at the same time, improve the network intrusion detection speed.