Applied Research Of Data Mining In The Police Integrated Information System

As the informationize of police work continue to push forward, the application of a large number of police information systems, in which we accumulate a large number of cases and suspects and other police data. Vast amounts of police data stored in different data sources, which has a lot of data quality problems, is difficult to comprehensive analysis; due to the different emphases of data collection and management in different business systems, the information reflected in one kind of data source is incomplete, analysis in a separate data source brought misjudgments and leads are lost, etc. while the data are incomplete or missing; a lot of analysis depend on manual, resulting in a large number of energy of analysts focus on repetitive work, like access to intelligence clues, sorting, cleaning and comparision, but not able to concentrate on the analysis at the operational level; high-value information is often very small, and usually,the higher its value, more time-strong, and the current analytical ways led to a longer time to obtain information. When the leadership and decision-makers need to analysis, which needed a lot of manual reports and statistic, not only cause a heavy workload but also cause slowly analysis process, resulting in accurate and untimely guidance.Data mining is a technology which combines database, artificial intelligence, machine learning, statistics and other fields theory and technology, extracting implicit, unknown, and potentially non-trivial value of the information or model from large database or data warehouse, has been widely used in scientific research, banking, telecommunications, insurance, transportation, retail, and many areas. However, due to the sensitivity and confidentiality of police data, the research and application of data mining technology in the Police Integrated Information System on lags behind other areas. Therefore, large number of police data can not be effectively applied, can not be transformed into knowledge to provide support for the police decision.In this paper, the research is on the background of the Police Integrated Information System, analyzed the data sources and data structure of police data, used CDC, Kettle data extraction tools, using data extraction, cleaning, integration, conversion, reduction ways on various police data, established police data warehouse; In the process of data mining analysis, data cube is introduced, analyzed several model of data cube, and improved the processing speed of mining analysis by partial materializing police data cube; The use of association rule mining, decision tree analysis, related cases mining,etc., and other mining algorithms, combined with the police data for applied research. Police data on traveller data for mining frequent itemsets, association rules generated according to the target attributes; generated decision tree using traveller data, and then simplify the decision tree by pruning and other methods, to generate decision rules; Converted and integrated the cases data by similarity matrix, for mining related-cases; Finally, design and implementation decision analysis and multidimensional analysis of police data in the Police Integrated Information System.