摘要
[目的/意义]对多源异构时空数据进行同现模式挖掘可以发现涉恐人员、涉恐物资、涉恐活动在邻近地点同时段共同出现的规律,为反恐工作提供包含时间特征和空间特征的情报信息。[方法/过程]在同位模式挖掘的基础上提出预先确定目标情报类别从而确定涉恐人员分类方式,使用概念层次树对涉恐物资和涉恐活动进行数据预处理,使得目标同现模式覆盖更多情报信息。[结果/结论]该方法与同位模式挖掘、频繁时间序列模式挖掘以及各种时空轨迹模式挖掘可以相互补充,覆盖多种不同的反恐情报信息,完善反恐预警机制,为反恐决策提供客观依据。
[Purpose/Significance]Co-occurrence pattern mining of multi-source heterogeneous spatiotemporal data could discover the law of co-occurrence of terror-related persons,terror-related materials and terror-related events in adjacent locations at the same time,and provide intelligence information with temporal and spatial characteristics for counter-terrorism work.[Method/Process]On the basis of co-location pattern mining,this new method pre-determines the intelligence category of the target in order to determine the classification method of terror-related persons,and uses the concept hierarchy tree to preprocess the data of the materials and activities,so that the final co-occurrence patterns could cover more intelligence information.[Result/Conclusion]This method,together with co-location pattern mining,frequent time series pattern mining and various spatial-temporal trajectory pattern mining,can complement each other,cover a variety of different counter-terrorism intelligence,improve the counter-terrorism early warning mechanism,and provide objective basis for counter-terrorism decision-making.
作者
李勇男
Li Yongnan(School of National Security,People's Public Security University of China,Beijing 100038)
出处
《情报杂志》
CSSCI
北大核心
2021年第2期90-94,68,共6页
Journal of Intelligence
基金
中央高校基本科研业务费项目“反恐情报异常时空轨迹数据分析方法研究”(编号:2020JKF303)
中国人民公安大学国家安全高精尖学科资助项目(编号:2020GDLW004)的研究成果之一。
关键词
数据挖掘
反恐情报
时空同现模式
时空数据
同位模式
兴趣度
data mining
counter-terrorism intelligence
spatiotemporal co-occurrence pattern
spatiotemporal data
co-location pattern
interestingness
