论文提出了基于密度的异常挖掘新方法,并将其应用于入侵检测系统引擎设计中,构建了相应的网络入侵检测系统。该系统通过挖掘异常数据的高效性,可及时发现新的未知入侵行为,用以更新入侵规则库。基于该规则库,系统采用BM模式匹配算法进行实时入侵检测。论文运用形式化语言对入侵检测系统各子模块进行结构化分析与描述。 The dissertation proposes a new density-based anomaly mining method, and applies it to the engine design of intrusion detection system, and builds a corresponding network intrusion detection system. By mining the efficiency of anomalous data, the system can discover new unknown intrusion behaviors in time to update the intrusion rule base. Based on the rule base, the system uses BM pattern matching algorithm for real-time intrusion detection. The thesis uses the formal language to carry on the structural analysis and the description to each sub-module of intrusion detection system.