基于图的数据挖掘在入侵检测系统中的应用

Graph-based data mining for intrusion detection system

网络入侵检测系统(IDS)是保障网络安全的有效手段,但目前的入侵检测系统仍不能有效识别新型攻击。根据国内外最新的图数据挖掘理论,设计一个特征子图挖掘算法,并将其应用到入侵检测系统中。该算法挖掘出正常的特征子结构,与之偏离的子结构为异常结构。实验结果表明,该系统在识别新型攻击上具有较高检测率。

Intrusion Detection Systems (IDS) are developing very rapid in recent years, while the networks are being used widely. But most of traditional IDS can't detecting new attacks. Graph-based data mining is a subject that occurred in the past few years. Based on the theory of graph-based data mining, an algorithm of mining the substructures of a graph was designed, and it was applitd into IDS. It can mine normal pattern from graph data. The result of experiment shows that it can detect new attacks efficiently.