文献类型：期刊文献

英文题名：An immune system-inspired paradigm for anomaly detection

作者：Peng, Lingxi[1,2];Li, Tao[1];Liu, Xiaojie[1];Chen, Yuefeng[2];Liu, Caiming[1];Liu, Sunjun[1]

机构：[1]Sichuan Univ, Coll Comp Sci, Chengdu 610065, Peoples R China;[2]Guangdong Ocean Univ, Sch Informat, Zhanjiang 524025, Peoples R China

年份：2007

卷号：4

期号：7-8

起止页码：1394

外文期刊名：JOURNAL OF COMPUTATIONAL AND THEORETICAL NANOSCIENCE

收录：SCI-EXPANDED(收录号：WOS:000252406200031)、、EI(收录号：20080711090454)、Scopus(收录号：2-s2.0-38849138602)、WOS

语种：英文

外文关键词：artificial immune system; anomaly detection; machine learning; classification

外文摘要：Network anomaly detection is a vibrant research area. Researchers have approached this problem using various techniques such as artificial intelligence, machine learning, and state machine modeling. In this paper, we first review these anomaly detection methods and then describe in detail an adaptive anomaly detection paradigm based on an artificial immune system. The paradigm is referred as AIPAD. The implementation of paradigm includes: first, an initial antibody set is created; then, through the learning of each training antigen, the artificial immune system is evolved and updated by optimal antibodies; and finally, detection is accomplished by the majority vote of k nearest to the antibody in the network. The experiments use the famous Hepatitis Domain dataset in our study, which is taken from the UCI machine learning database. The obtained detection accuracy of AIPAD is 93.5%, which is very promising with regard to the other classification applications in the literature for this problem. In addition to its nonlinear classification properties, AIPAD possesses biological immune system properties such as clonal selection, immune network, and immune memory, which can be used to pattern recognition, classification, etc.