Computer Engineering and Applications ›› 2021, Vol. 57 ›› Issue (15): 156-162.DOI: 10.3778/j.issn.1002-8331.2004-0251

Previous Articles     Next Articles

Dendritic Cell Model Using Singular Value Decomposition and Information Gain

YANG Xinmin, DONG Hongbin, TAN Chengyu, ZHOU Wen   

  1. 1.School of Cyber Science and Engineering, Wuhan University, Wuhan 430072, China
    2.School of Computer Science, Wuhan University, Wuhan 430072, China
  • Online:2021-08-01 Published:2021-07-26



  1. 1.武汉大学 国家网络安全学院,武汉 430072
    2.武汉大学 计算机学院,武汉 430072


Aiming at the problem that the signal extraction process of the Dendritic Cell Algorithm(DCA) is affected by artificial experience and its ability to detect anomalies in disordered data is not strong, SIDCA, a Singular Value Decomposition(SVD) and Information Gain(IG) dendritic cell model is proposed. SIDCA uses the SVD method to obtain the most relevant feature subset, and then uses the information gain to extract the most relevant feature in the most relevant feature subset to realize adaptive signal extraction and reduce the disorder of data to the algorithm confused. Experimental comparisons with classic DCA and deterministic DCA(dDCA)show that SIDCA has higher accuracy and lower false alarm rate on ordered and unordered data sets.

Key words: dendritic cell algorithm, Singular Value Decomposition(SVD), Information Gain(IG), computer immune, danger theory


针对树突状细胞算法(Dendritic Cell Algorithm,DCA)存在的信号提取过程中受人工经验影响和对无序数据的异常检测能力不强的问题,提出了采用奇异值分解(Singular Value Decomposition,SVD)和信息增益(Information Gain,IG)的树突状细胞模型——SIDCA。SIDCA用奇异值分解(SVD)方法得到最相关特征子集,再使用信息增益提取最相关特征子集中的最相关特征,实现自适应提取信号,降低无序数据给算法带来的混淆。实验结果表明,与经典DCA和确定性DCA(Deterministic Dendritic Cell Algorithm,dDCA)相比,SIDCA在有序和无序数据集上均有更高的准确率和更低的误报率。

关键词: 树突状细胞算法, 奇异值分解, 信息增益, 计算机免疫, 危险理论