关键词: 树突状细胞算法, 粗糙集, 入侵检测, 属性约简, 属性重要度

Abstract: It is necessary for Dendritic Cell Algorithm（DCA） to reduce the monitoring data on network, which can alleviate the burden of intrusion detection system and improve the efficiency of detection. This paper proposes an intrusion detection method by integrating the DCA with attribute reduction in rough set theory. By using the concept of attribute significance in rough set theory, it reduces the attribute of dataset and produces the input signals of DCA, then detects the intrusion by DCA. The improved algorithm is tested based on the KDD CUP 99 dataset. The results show that the improved algorithm can reduce the false positive rate when the detection rate is guaranteed. The improved algorithm realizes the auto-extracting of intrusion features and reduces the features needed exactly, accelerates the speed of algorithm. The improved algorithm performs well in combination property.

Key words: dendritic cell algorithm, rough set, intrusion detection, attribute reduction, attribute significance