基于词性的文本挖掘算法在IDS日志中的应用  

A Text Mining Algorithm Based on Part of Speech Used in IDS Logs

作　　者：胡军光 刘力[2] 车奇 

机构地区：[1]空军驻深圳地区军事代表室,深圳518026 [2]南京航空航天大学信息科学与技术学院,南京210016

出　　处：《计算机与数字工程》2010年第2期90-93,共4页Computer & Digital Engineering

摘　　要：提出一种以词性为参考值的文本挖掘算法,能有效挖掘与种子词有关的关联规则。基于Bootstrapping算法思想,既减少了预处理阶段对于词根还原的依赖,能处理日志中出现的中文词汇。增加了对于日志文本上下的理解,提高了关联规则的有效性,并应用与IDS日志挖掘之中,有效改善挖掘效率,为规则库提供关联规则。We made a text mining algorithm using part of speech （POS） as its argument, which can effectively mine the seed-related rules. Based on the idea of Bootstrapping algorithm, it can reduce the dependence of root-restoring on the pre- processing stage, process Chinese vocabulary appear in the log, increase the understanding of context, enhance the effectiveness of rule-relating. When applied in IDS log mining, it will significantly improve the mining efficiency and provide rule li- brary with rules.

关 键 词：词性 ICTCLAS BOOTSTRAPPING 入侵检测系统 数据挖掘 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]