终身学习的否定选择算法  

Negative selection algorithm for life-long learning

作　　者：张福勇[1] 齐德昱[1] 胡镜林[1] 

机构地区：[1]华南理工大学计算机系统研究所,广州510006

出　　处：《沈阳工业大学学报》2012年第3期293-297,共5页Journal of Shenyang University of Technology

基　　金：国家技术创新基金资助项目(08C26214411198);粤港关键领域重点突破基金资助项目(2008A011400010)

摘　　要：针对否定选择算法(NSA)仅模拟了人类适应性免疫系统的中枢耐受过程,而未模拟其外周耐受过程,不具备终身学习能力的问题,提出了具有终身学习能力的否定选择算法(LNSA).LN-SA的学习过程模拟淋巴细胞经历中枢耐受和外周耐受的成熟过程,通过外周耐受实现终身学习,从而提高了抗体区分自体与非自体的能力.与NSA和树突细胞算法(DCA)的比较结果表明,LNSA可有效降低检测误报率,并具有与NSA同样高的检测率.In order to solve the problem that the negative selection algorithm（NSA） only simulates the central tolerance process for human adaptive immune system,can＇t simulate the peripheral tolerance process,and has no capacity of life-long learning,the life-long learning negative selection algorithm（LNSA） was proposed.The learning process of LNSA simulates the maturation process of both central and peripheral tolerances suffered by lymphocyte,and the life-long learning is realized with the peripheral tolerance.Therefore,the capability of distinguishing self and nonself for antibodies gets enhanced.Compared with NSA and dendritic cell algorithm（DCA）,the LNSA can effectively reduce the detection false positive rate,and exhibits the high detection rate like NSA.

关 键 词：人工免疫系统 恶意代码检测 否定选择算法 树突细胞算法 终身学习 适应性免疫 n元 I/O请求包 

分 类 号：TP309.5[自动化与计算机技术—计算机系统结构]