Research On Real-time Recommendation Of Crime Clues Based On Parallel Processing Of CEP

Throughout history,the crime has been accompanied by the date of the country,which exists in various historical periods and various social types,and has never disappeared.The development of network technology provides more method for criminal investigation.Relying on the network to carry out the crime clues collection will become a new method of crime investigation.With the rapid development of Internet technology,the information data of the network have already increased exponentially.It is an urgent problem to be solved that we should quickly find relevant clues and recommended to the relevant departments of criminal investigation from the mass data.Complex Event Processing(CEP)is a process,which abstract the external data stream into an event,and then use the filtering,aggregation and other operations of the event to detect the event patterns defined in advance and respond to it.Its greatest advantage is the process the real-time data,which allows us to get the most useful information in the shortest time.The parallel complex event processing can effectively solve the problem of centralized processing,and it has a great advantage in the processing the mass data.On the basis of the collection of the original crime clues,we introduce parallel complex event processing technology into crime clues recommendation system for the relevant departments,which provide real-time and efficient guarantee.In this paper,we study the parallel complex event processing technique,on the basis of the existing network crime clues collection system,combined with the advantages of complex event processing in real-time processing,to carry out research and implementation of real-time recommendation system of crime clues based on the parallel CEP.Keywords semantic trees are built by analyzing the relationship between duty crimes.Node weight is the node number in the tree.In this paper,we calculate the number of matching keywords and total weights according to the semantic relations of keywords.Recommended rules are based on total weights.On the basis of analysis of the event patterns,some key steps in the patterns research and design are given.We design and implement of the core module of the system.Experimental results show that the accuracy of this method is higher than traditional method;the system has a certain advantage in processing mass data and high real-time requirements.