Research And Application In Criminal Interrogation Assistant Decision Algorithms Based On Cases

Criminal interrogation is the most important and complex task of police system. Through asking suspects for the criminal reality and correlative questions is revealing the truth, proving the criminal reality, finding suspects. The fundamental research object of the interrogation work is suspects' psychological activity and behavior performance in the different interrogational stage. This systematized knowledge hides in the huge crime case information deposited in the police system. Current interrogation process costs massive manpower and resources, has the phenomenon of wrong justice, and sometimes causes to delay the investigation opportunity situation because of the process time is long. Moreover, at present in this domain, there are many scattered service data, which has been distributed in different departments; Department's data is processed independently and the relations need to be manually established, so the efficiency is low and data lacks the intellectualized statistics and analysis.This paper takes the research background of Dalian police project - Criminal Interrogation Assistant Decision Support System. This paper illustrates data mining technology in the criminal interrogation cases, including discussing from three different aspects.Firstly, this paper research a lot of technologies about Rough set and applies Rough set to attribute discrete of criminal cases. This method sufficiently thinks correlative attribute's influences and availably reduces unreasonable and redundant partition.After attribute discrete, the paper gives the Bayesian Network model about partial criminal interrogation cases based on the correlative technologies of Bayesian Network, including crime suspect psychology characteristic classification model and interrogation strategy model.Finally, this paper analysis a lot of criminal cases by using the association rules, finishes data mining application in criminal cases, has the potential rules about data records and builds the association models. It applies to finding the rules from a lot of correlative cases and puts forward feasible interrogation projects to people. It improves the interrogation accuracy and efficiency. Because of the proliferation of criminal cases, the paper brings forward an incremental updating algorithm for mining association rules.The paper applies different technologies in knowledge discovery to the data mining in the criminal interrogation cases and gets reasonable Data Mining model. The model is composed of opening problem definition, data mining and consequence. It has universality and expansibility toanalyze all kind of questions. The results of experiment have obtains the police professional's approval.