网络入侵检测系统中的频繁模式挖掘

Mining Frequent Patterns in Network Intrusion Detection

为了解决网络入侵检测领域使用Apriori算法挖掘频繁模式效率不高、精度不够的问题,在FP-growth的基础上提出一种新的基于分割原理的PFP-growth算法。该算法采用分而治之的方法,既有效利用了FP-tree特性,又减轻了系统挖掘大容量数据库的负荷,使挖掘效率有了明显提高。另外设计了一种新的最小支持度设置法,使挖掘的频繁模式更精确。

In the network intrusion detection, Apriori algorithm is used to extract relative rules, but its processing precision and efficiency are not satisfactory. In order to resolve the problem, based on FP-growth, this paper proposes a new algorithm named PFP-growth, this algorithm applies an idea of divide and rule, makes good use of FP-tree, and eases the load of system when mining a large database, which makes its velocity improved obviously. Besides we design a new method to set min-support, which makes frequent patterns mined much precise.