基于数据流聚类挖掘算法的入侵检测系统研究

Research on Intrusion Detection System Based on Data Stream Clustering Mining Algorithm

为解决传统入侵检测实时性不足的问题,针对当前网络安全中处理速度快、防御能力强、实时性能高等特点,研究基于数据流挖掘与入侵检测相融合的网络安全防御技术,建立新的检测模型,设计了快速剔除孤立点的算法,提出一种改进的基于衰减滑动窗口密度的数据流聚类挖掘(ASWDStream)算法,并对该算法及其在入侵检测系统中的应用效能进行了验证。仿真结果表明该算法具有较低的运行环境要求和较高的聚类准确性,入侵检测系统表现出较高的检测率和实用性。

To solve the problem of real-time performance in the conventional intrusion detection system,the characteristics of current network security were analyzed,such as high processing speed,strong defense capability,real-time performance.Then this paper studied the network security defense technology based on data stream mining and intrusion detection,and a new detection model was established.Then this paper designs a way of removing isolated points,and an improved algorithm of attenuation sliding window density-based data stream clustering mining(ASWDStream)was proposed.And the application effectiveness of the algorithm in intrusion detection system was verified.The simulation results show that the algorithm has a lower requirement for operating environment and higher clustering accuracy.The intrusion detection system shows a higher detection rate and availability.