基于离线电子对抗历史数据的信息特征关联分析

Information feature association analysis based on historical data of off-line electronic countermeasure

针对复杂电磁环境电子对抗信号分析要求高、情报处理数据量大、多维信息分析提取困难等问题,基于电子对抗情报系统海量历史侦察数据特点和已有电磁目标知识库,研究跨行业数据挖掘标准流程(CRISP-DM)大数据处理方法,设计了一种离线电子对抗电磁信号历史数据处理分析平台。通过探索挖掘电磁目标参数、目标时间规律、多目标关联规则等情报分析的技术路径,实现了聚类、时间序列和关联规则挖掘等分析方法在电子对抗电磁信号历史数据处理中的应用;获取了未知电子目标聚类、目标数量规模预测和多目标关联共现规律等信息特征。结果表明,电磁目标参数特征和关联规则明显,目标时间特性拟合相关度达0.825,为后续进一步研究及实践应用提供了参考。

In view of the high requirements of electronic countermeasures signal analysis,the large amount of intelligence processing data,and the difficulty of multi-dimensional information analysis and extraction in complex electromagnetic environment,the big data processing method of cross-industry data mining standard process,namely Cross-Industry Standard Process for Data Mining(CRISP-DM),is studied,and an processing and analysis platform for historical data of offline electronic countermeasures electromagnetic signal is designed based on the characteristics of the massive historical reconnaissance data of the electronic countermeasures intelligence system and the existing electromagnetic target knowledge base.By exploring the technical path of mining electromagnetic target parameters,time rules of the target,multi-target association rules and other intelligence analysis,the application of analysis methods such as clustering,the mining of time series and association rules in the processing of electronic countermeasure electromagnetic signal historical data is realized.The information features of unknown electronic target such as clustering,target quantity and scale prediction,multi-target association and cooccurrence rule are analyzed and obtained.The results show that the characteristics and correlation laws of electromagnetic target parameters are obvious,and the fitting correlation degree of target time characteristics reaches 0.825.This work lays a foundation for further research and practical application.