Abstract: At present, most of the community correctional institutions have adopted the traditional methods such as manual management, lacking of effective supervision of the cross-regional movement and daily dynamic community corrections, and there are the lack of supervisors and real-time supervision problems. Based on the above problems, this paper adopts the algorithm of data association analysis, abnormal behavior analysis and decision tree analysis, making multi-dimensional data mining on the location information, case information and related behavior information of community prisoners. Community correction managers use the relevant results of the analysis to manage some work more effectively. These data analysis technologies support the development of community correction management work well.

Key words: community correction, association analysis, abnormal behavior analysis, decision tree analysis, data mining

摘要： 目前社区矫正机构对社区矫正人员的管理大多还采用人工管理等传统方法，对社区矫正人员的跨区域流动、日常动态等情况缺乏有效的监管，存在监管人员不足、不能实现实时监管等问题。基于以上问题，采取数据关联分析、异常行为分析，决策树分析等算法模型，对社区服刑人员的定位信息、案件信息、相关行为信息等进行多维度的数据挖掘。社区矫正管理人员利用数据挖掘得出的相关分析结果对矫正人员进行有效的管理。这些数据分析技术很好地支撑了社区矫正管理工作的开展。

关键词: 社区矫正, 关联分析, 异常行为分析, 决策树分析, 数据挖掘