计算机技术与发展2019，Vol.29Issue(3)：1-5,5.DOI:10.3969/j.issn.1673-629X.2019.03.001
人工免疫算法在AETA异常检测中的应用研究
Research on Application of Artificial Immune Algorithm in Discriminating Abnormal Data of AETA Equipment
摘要
Abstract
AETA multi-component seismic monitoring system has been deployed in more than 200 sets in Sichuan, Yunnan, Hebei, Guangdong, Tibet and Taiwan. A set of AETA equipment generates 7 G of data every month, and it takes a lot of time and effort to view the data of 200 sets of equipment, and it is easy to miss detection by manual judgment. Using the idea of negative selection in the artificial immune algorithm, the corresponding antibody library is obtained by learning the normal data. According to the matching degree between the data to be tested and the antibodies in the antibody library, the abnormal detection of equipment data is carried out. In addition, combined with the data characteristics of AETA equipment, the affinity function in the artificial immune algorithm is improved to reduce the amount of calculation. First, the normal data of low-frequency electromagnetic data of 8 typical stations are selected from the AETA equipment for training to obtain the antibody library, and then anomaly detection is performed on the measured data. By this method, it is possible to effectively detect three types of abnormal data in the system. The detection accuracy of the three abnormalities are 87.20％, 83.33％ and 55.56％ respectively. A preliminary determination of abnormal data can be achieved.关键词
人工免疫算法/异常检测/地震数据分析/数据挖掘Key words
AIS/abnormal detection/earthquake data analysis/data mining分类
信息技术与安全科学引用本文复制引用
李柏杭,王新安,雍珊珊,徐伯星,黄继攀..人工免疫算法在AETA异常检测中的应用研究[J].计算机技术与发展,2019,29(3):1-5,5.基金项目
广东省省级科技计划项目(2014B090913001) （2014B090913001）
深圳市科技创新委员会项目(JCYJ20170412151159461) （JCYJ20170412151159461）
