入侵检测中数据挖掘模式的编码与检测研究  

STUDY ON ENCODING AND DETECTION FOR PATTERNS MINED BY KDD IN IDS

作　　者：熊家军[1] 张丽[2] 李庆华[1] 

机构地区：[1]华中科技大学计算机学院,湖北武汉430074 [2]空军雷达学院计算机系,湖北武汉430019

出　　处：《计算机应用与软件》2005年第11期11-13,95,共4页Computer Applications and Software

基　　金：国家自然科学基金资助项目(编号:60273075)。

摘　　要：数据挖掘产生的模式如何被处理,对IDS系统的实时性和有效性至关重要。该文提出了一种模式编码算法,对模式进行量化处理,从而高效地检测入侵。同时,文中还从理论上论证了模式编码量化检测方法与基于相似性理论的模式检测方法的等效性,证明任何用相似性理论实现的入侵模式检测算法均可用编码量化处理来等效实现。How to use the patterns mined by KDD algorithms is very important to real-time and effectiveness of IDS. This paper put forward an algorithm of pattern encoding, which can expresses patterns in digits, thus the intrusion pattern can be detected efficiently. At the same time,it also was proved that the patterns detection methods based on pattern encoding is equal to the methods with similarity theory, as a resuit, any intrusion pattern detection algorithms based on similarity theory can be realized by pattern encoding.

关 键 词：入侵检测 数据挖掘模式 关联规则编码 模式编码 频繁情节规则编码 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]