摘要
针对传统模糊C-均值聚类算法(FCM)收敛速度慢、易陷入局部最优的缺点,用带交叉操作的改进微粒群算法来弥补FCM算法的不足,提出一种优化的模糊C-均值聚类算法(OFCM)。模拟仿真实验表明该算法具有较快的收敛速度和很好的全局搜索能力,解决了FCM算法在入侵检测中稳定性差、检测精度低的问题。新算法在网络安全方面有很好可行性和实用性。
In view of the faults of the traditional fuzzy C-means clustering algorithm in convergence speed and easy fall into local optimum, the improved particle swarm optimization algorithm with cross-operation is used to make up for the deficiency of the fuzzy C-means （FCM） algorithm, thus an optimized fuzzy C-means clustering algorithm is proposed. Simulation experiment results show that the optimized fuzzy C-means （OFCM） algorithm is better than FCM in global searching capability and convergence speed. The method overcome the shortages of FCM, such as poor stability and low intrusion detection precisiorn The new algorithm has good feasibility and practicality in network security.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第11期4100-4104,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(61003035)
关键词
入侵检测
模糊C-均值聚类
交叉操作
隶属度
适应值
intrusion detection
fuzzy C-means clustering
crossover operation
membership degree
fitness value
