网格空间资料探勘应用于犯罪预测之研究

Crime Prediction Using Grid Spatial Analysis and Data Mining Techniques

台湾警察机关自1970年代起陆续将各项警政资料,包括治安、交通、犯罪侦查等相关数据建立信息系统,目前已累积数量庞大的治安大数据。善用信息科技强大的数据处理、统计分析及探勘功能,挖掘隐藏于大数据的知识,以提供决策所需的辅助信息,不但有助于掌握问题及研拟因应作为,而且经由趋势分析与模式归纳学习,更可以进一步预测未来,作为规划因应方案的参考。此项研究搜集整理台湾新北市刑案记录数据,运用网格分析法,定义网格毒品犯罪形态特征及设计毒品犯罪预测向量,并以数据探勘软件套件进行犯罪预测实验。结果显示,以网格分析方法及大数据探勘预测犯罪之平均效能,明显优于传统运用经验法则的犯罪预测方法。

Since 1970s, Taiwan police agencies have been using information systems to process policing data, such as law enforcement data, traffic management data, and crime investigation records, etc. Today, Taiwan police agencies have accumulated a huge amount of policing big data. The advancement of data mining techniques has made it possible to explore and dig out hidden knowledge from big data for supporting management tasks and decision making processes. In this research we used both grid spatial analysis and data mining techniques to predict drug crime in Xinbei Municipal City. The experiment results show that the effectiveness of the heuristic prediction method. our prediction method is much better than that of