机构地区: 中国人民公安大学侦查与反恐怖学院,北京100038
出 处: 《情报科学》 2017年第9期141-145,152,共6页
摘 要: 【目的/意义】利用数据挖掘技术在海量数据中快速、准确、有效的发现涉恐线索并及时处置是反恐工作的重要手段之一。【方法/过程】本文研究如何利用频繁模式树对涉恐基础数据进行挖掘,提取涉恐特征的频繁项集。首先通过对数据库中的涉恐人员信息进行涉恐特征计数排序并建立频繁模式树,然后在树结构中递归遍历发现满足最小支持度阈值的频繁项集。【结果/结论】文中的方法可以快速发现大量基础数据中的涉恐关联属性,有利于在系统中自动搜索重点涉恐人员,为反恐预警系统提供决策参考。通过与其他产生关联规则的方法结合使用,还可以发现暴恐活动中不同因素的因果关系。 【Purpose/significance】It is one of the most important methods to accurately find valid terrorist activity clues in time from mass data by using data mining technologies.【Method/process】This paper proposes how to extract frequent itemsets of terror related features by using frequent pattern-growth algorithm.These features are sorted by counting the numbers to construct frequent pattern tree.The frequent itemsets are generated by traversing the tree structure recursively.【Result/conclusion】This method could quickly find the associated terror related features.These relationships are conducive to search the key person automatically from the database in the system and provide decision references for terror threat warning.With the help of association rule mining,this method is also useful to find the potential terrorist activities.