关键词: 入侵检测, 核心微簇, 缓冲微簇, 聚类纯度

Abstract: The capacity of dealing with data streams has become a challenge for intrusion detection system. A DC-stream algorithm is proposed in view of this situation. Adopting both online and offline clustering, the DC-stream algorithm designs a buffer type anomaly detection mechanism, which can not only ensure the efficiency and accuracy of the data stream clustering, but also filter the noise data. The experimental result shows that the algorithm can detect intrusion behaviors in the mass network data stream timely and effectively, and has strong anti-
interference ability.

Key words: intrusion detection, core micro cluster, buffer micro cluster, cluster purity