基于分布式数据入侵检测模型研究

Research on updating algorithm of association rules based on distributed intrusion detection

在数据挖掘技术应用于入侵检测的研究中,分布式环境下的全局频繁项目集的更新算法尚不多见。为改善入侵数据增加后,更新算法的效率问题,提出一种基于分布式入侵检测的关联规则快速更新算法TDUA(Trivial Distributed Update of Association rules)。算法引入强频繁项目集概念,有效地修剪候选项目集数目,采用共享模式的体系结构,具有通信代价较小,挖掘效率较高的特点。并从理论与实验两个方面验证了算法适应在高速大流量的分布式网络环境下快速准确地建立入侵检测模型。

In the research of intrusion dection which data mining technology was used,it was rared to see updating algorithm of the global frequent item sets in distrbuted database.This paper proposed a updating algorithm TDUA（trivial distributed update of association rules） in order to improve the efficiency of updaing algorithm in the case that invading data increased.The algorithm introduced the concept of frequent item sets strong and effectively prunes the number of candidate itemsets.It used the shared model architecture and had the feature of a less cost for communication and more efficient for mining.From both theoretical and experimental aspects,the algorithm is verified to adapt to build intrusion detection model quickly and accurately in high-speed and high-volume distributed network environment.