At present, the construction of information in the police system has already reached a considerable scale, but the major applications remain in the simple and traditional functions such as enquiry, statistics, and lacks in the deep-level analysis. The public security departments urgently need automatic tools to mine the patterns and trends which hidden in the crime data.Data mining of criminal characteristics system is based on the database and achievement in Public Security Bureau and focus to criminal cases, which study the data mining technology based on police business information and design and develop the corresponding data mining tools.The thesis describes the overall design of the system, including the demand analysis of functions, design objectives, system architecture and system interface, the discussion focused on the data pre-processing platform and case clustering module.Because the original database of Public Security Bureau is built for day-to-day operation and is not suitable for data mining. Through data extraction, data cleaning, data conversion and data loading the data pre-processing platform integrate the original data into the data warehouse. The thesis discussed the design and implementation of the platform in detailed and highlighted the accuracy and safety in conversion.Case clustering is one of the core functions of the system. The k-modes algorithm is relatively efficient in processing large data sets and can work on data with categorical attributes, so it is used as the cluster analysis algorithm in the system. The thesis discusses the implementation of the algorithm and also analyzes the results of implementation and performance. |