Analysis of decision tree classification algorithm based on attribute reduction and application in criminal behavior

In this paper, the attribute reduction strategy is syncretized into classification algorithm of the decision tree and applied to criminal behavior analysis. Finding implicit knowledge in the criminal database by this method can utilized to assist making decision for police quickly and accurately. The classification algorithm of the decision tree based on rough set is proposed for multi-attribute data table. The scale of decision tree and branches is minished and the reliability is improved via attribute reduction. Successful application in the analysis of criminal behavior shows the feasibility of the algorithm.