【摘要】 为克服模糊C均值(FCM)算法对初始化极为敏感且容易陷入局部最优的缺点,将遗传算法和改进的模糊C均值聚类算法相结合,并以检测率和误检测率作为入侵检测算法性能评价的指标,对FCM、改进的FCM、基于遗传的改进FCM3种聚类算法的入侵检测性能进行仿真分析。仿真实验表明,结合遗传和FCM两种算法的混合算法能够实现优势互补。由于该算法结合了遗传算法,使整个算法的复杂度增加。从入侵检测看,通过增加处理时间而提高了入侵检测率。更多还原

【Abstract】 In order to overcome the shortcomings of the algorithm of Fuzzy C-Means(FCM) ,namely,the extreme sensitivity to the initial data and the liability to come to local optimum points,this paper proposes a combination of the genetic algorithms and the improved fuzzy C average value methods and the use of the examination rate and the mistaken examination rate as the evaluation indices for the invasion examination algorithm performance. The simulation analysis is carried out for the invasion examination performance of three kinds of algorithms related to FCM,improved FCM and improved FCM based on the heredity. The result indicates that the method,based on the genetic improved FCM algorithm(GIFCM) enjoys higher examination rate and lower mistaken examination rate. This algorithm can find an effective application in exceptionally invasion examination.更多还原