Research And Implementation Of Serial Case Analysis Based On Clustering

With the continuous development of social economy and the standard of people's living,the trend of crime professionalism,grouping and fleeing is becoming more and more prominent.The number of crime cases is increasing day by day.Larceny cases increased rapidly,and the crime cases by the same group occupied a considerable proportion.It's particularly important to investigate a series of cases.Public security has accumulated a lot of criminal information data and detection experience over the years.It's very important to analyze and mine the data so that we can find the law and the trend of crime.And we also can find the relationship between the criminals.This is the main task of the public security.In this paper,the research is based on the data of some cases of financial crimes in Changning District recent years.Based on the features we can determine which cases are belong to the same series.Firstly,we can observe the characteristics of each case,and confirm the scope of information extraction so that we can prepare for the extraction.Secondly,extract the features that involved in the case description and then cluster all the data of cases and analysis.Through the results of cluster analysis that we can see the cases of the same cluster as a series of cases and then we can get the result of analysis series of cases.Finally,combined with the needs of system,I designed and implemented a series cases analysis system.The system is mainly divided into three modules: case feature extraction,cluster analysis and display module.Case feature extraction module is mainly used to extract the characteristics of the cases in unstructured data.Firstly,segment the text by ansj algorithm and CRF algorithm to sign the semantic annotations.Then extract the crime features by pattern match algorithm.Cluster analysis module has two process: preconditioning the feature data and cluster computing by FCM algorithm.Chose a reasonable number of categories and weight coefficient M value.Finally,clustered and analyzed the data set.Display module is used to show the results with the form of graphs and tables.The website is based on JavaEE and ECharts.Finally,through lots of experiments,it verified that the system in the study can satisfy the requirements,provided more intuitive and reliable case features and relations to the analysis.In this way,they can infer some clues and information more easily which has a certain application value.