The Research On Police Crime Analysis Using Data-mining Technology

Data mining technologies have been more in-depth researched in many fields,but just have elementary been researched and applied to the public security work,many public security information systems stills stay at a low level, lack the synthetic exploiture and application,lack the intelligence analysis estimation,lack the scientific decision prevision and alarm, lack the the ability of process from microcosmic to macroscopical in data,the means of discover a problem from microcosmic to macroscopical in data.This thesis focses on how to apply data mining technologies to mining and extract the associable rule and trend in the a great deal of data information,in order to enhance the executive efficiency and quickly respond ability,prevention and fight crime in time,and provide the support service for the decision-making in the police affair.The thesis primary work as follows:The thesis contrapose the conventional apriori algorithm to exist the problem in public security crime analyze: ignore the problem of different item and different importance, put forward the reformative weighted association rules algorithm WARMA. the reformative algorithm design the weighted association rules model, Expel??o become impossible the weighted frequent k-item of the candidate subset through using k-support expectation, resolve the problem that subset of the weighted frequent item can not be the weighted frequent. experiment and analyse indicate: algorithm WARMA has the less candidate subset and the shorter runtime,can more effectively discover important crime.The thesis contrapose the conventional apriori algorithm to exist the problem in public security crime analyze:can not duly observantly discover the new crime, put forward the reformative association rules incremental updating algorithm SFUP base on the algorithm FUP. algorithm SFUP scale the recognition circs for new item through using the the sensitivity parameter, improve the ecumenical incremental updating algorithm come into being frequent item process. experiment and analyse indicate: the reformative algorithm more big data quantity more advantage, can effectively discover new crime,and showed good space and time capability in the process of mining, consumedly heighten the all mining efficiency.The thesis contrapose the conventional ID3 algorithm to exist the problem in public security crime analyze: the one is the according to the minimum principle of the entropy value , the attribute to be listed by ID3 algorithm to should be judged the first , but not so important in realistic; Two is when the majority attribute data quantity are bigger, the separately attribute data quantity are lesser, easily appear the big data to flood the fraction, thus lose some correspond the important judgment. put forward the improved BID3 algorithm base on the transcendental parameter, the BID3 algorithm improve the attribute choice standard of the conventional ID3 algorithm, intensify the important attribute attention,through the increased transcendental parameter B,so as to the decision tree reduce to the dependence which takes the value more attribute, reduce to appear the phenomenon of a big data to cover up fraction possibly. experiment and analyse indicate: in the decision tree structure process of the more large-scale data gather, BID3 algorithm more advantage compare ID3 algorithm on the efficiency and the function.The thesis combine the actua public security crime analyze, expansion the design and experiment of criminal analyze prototype system base on the decision tree, put forward the function requirement and system flow chart, introduce the prototype system module structure and realization. experiment and analyse indicate:it is important realism meaning in the public security work that mining the amount of sea data ??he public security information database,using data mining technology, discover regulation trend, assist the decision-making in the police affair fleetly true.