ROCK聚类在反恐情报分析中的应用研究

发布时间：2018-02-25 22:24

本文关键词： ROCK聚类反恐情报数据挖掘雅卡尔系数相似性度量　出处：《情报杂志》2017年10期 　论文类型：期刊论文

【摘要】：[目的/意义]通过不断搜集和积累各种涉恐信息,利用数据挖掘等技术进行分析比对,提炼有价值的涉恐线索可以为反恐预警提供参考。ROCK聚类可以实现大量情报信息的快速简单分类。[方法/过程]首先要对涉恐数据进行二元化,其次根据雅卡尔系数计算原始数据间的相似性,然后不断重复计算公共链接数、计算相似性度量、合并簇三个操作,直到达到终止条件为止。[结果/结论]该方法与其他数据挖掘方法组合使用,可以减少待研判涉恐人员的数据量,只需逐条对比部分数据,显著提高涉恐情报分析的效率。
[Abstract]:[objective / significance] through the continuous collection and accumulation of various terrorism-related information, the use of techniques such as data mining for analysis and comparison, Extracting valuable clues about terrorism can provide a reference for early warning against terrorism. Rock clustering can achieve a quick and simple classification of a large amount of intelligence information. [methods / processes] first of all, the data involved in terrorism should be dualized. Secondly, the similarity of the original data is calculated according to the Yacar coefficient, and then the common link number is calculated repeatedly, the similarity measure is calculated, and the cluster is merged into three operations. [results / conclusion] the combination of this method and other data mining methods can reduce the amount of data of the persons involved in the study and only need to compare part of the data one by one, so as to improve the efficiency of intelligence analysis.
【作者单位】：中国人民公安大学侦查与反恐怖学院;
【基金】：2017年度教育部人文社会科学研究青年基金项目“基于数据挖掘的涉恐情报量化分析方法研究”(编号:17YJCZH098) 2015年度国家社会科学基金重大项目“当前我国反恐形势及对策研究”(编号:15ZDA034) 2014年度国家社会科学基金项目“反恐维稳背景下边疆地区维稳战略研究”(编号:14BZZ028)的研究成果之一
【分类号】：D631.4;G350

【相似文献】