摘要
聚类是一种根据数据相似度对海量数据进行"无监督"分类的信息处理技术。针对当前公安情报工作"数据丰富,知识贫乏"的现状,本文提出了一种基于聚类DBSCAN算法的情报分析程序,以某地市"两抢一盗"案件数据样本为例,测试了该程序在治安防范实践中的有效性,并与数字地图相结合,为预防多发性刑事案件提供了可视化的决策支持。
Clustering is an unsupervised data mining technology that is based on the similarity of records and used for the classification of data. Targeting at the ＂abundant data with poor knowledge＂ situation of policing intelligence work, currently, the paper offers an intelli gence analysis software based on DBSCAN algorithm, and takes ＂the robbery, grab and theft cases＂ data sample as example to test the ef fectiveness of the software on law enforcement. Moreover, the model is combined with the crime mapping, which can provide visual deci sion support for multiple and prolific criminal eases
出处
《情报杂志》
CSSCI
北大核心
2013年第8期27-30,共4页
Journal of Intelligence
基金
教育部人文社会科学研究青年基金项目"基于综合情报平台的重大突发事件预警防范机制研究"(编号:12YJC870003)
公安部公安软科学项目"基于智能情报平台的重大突发事件预警防范机制研究"(编号:2010LLYJGADX034)阶段性成果之一
