计算机免疫系统GECISM中识别规则的挖掘  

Mining of Recognition Rules in GECISM

作　　者：张寿华[1] 伊开[1] 王振夺[2] 任志利 刘振鹏[1] 

机构地区：[1]河北大学数学与计算机学院,河北保定071002 [2]北华航天工业学院计算机系,河北廊坊065000 [3]保定市第三中心医院微机室,河北保定071000

出　　处：《河北大学学报（自然科学版）》2007年第2期204-208,共5页Journal of Hebei University(Natural Science Edition)

基　　金：河北省科学技术研究与发展计划项目(0321350204213534)

摘　　要：GECISM(GEneral Computer Immune System Model)是基于规则匹配检测的计算机免疫系统,免疫识别规则对“自我”和“非我”特征的表征能力直接影响到GECISM的性能,所以挖掘高效免疫识别规则是GECISM的一个重要研究内容.改进后的Apriori算法以系统调用序列为数据源,从“自我”集和“非我”集中计算出频繁谓词,进而产生免疫识别规则.这些规则反映了“自我”和“非我”的内在特征,是GECISM进行“非我”检测的依据.GECISM is a rules matching based computer immune system whose performance directly relates to how much these recognition rules can represent the feature of self and nonself, so mining effective recognition rules is the key theme to implement GECISM system. Revised apriori algorithm finds frequent predicates from system call sequence which is the raw data, and then to generate recognition rules. These rules reflect some inherent feature of self and nonself, and they are the basic detection criterion of GECISM.

关 键 词：计算机免疫 识别规则 数据挖掘 APRIORI算法 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]