Violent Crime Factor Analysis Based On Particle Swarm Optimization And Rough Sets

The occurrence of many practical problems in real life is mostly the result of the interaction of the various subjects. So how to deal with cross-subject data is the tough questions in practice. Rough set theory, as a kind of new soft computing method, can be effective in the incomplete and inaccurate, inconsistent information data analysis and processing, to remove the redundant information in the premise of keeping the ability to distinguish, thus intelligent acquire valuable knowledge and information. Recently has been successfully applied to artificial intelligence, data mining, pattern recognition, fault diagnosis and many other fields. The results illustrate it is suitable for dealing with cross-subject data.This paper studies domestic and foreign research results of related technologies and analyzes the rough set attribute reduction algorithms. Focus on the implementation of various reduction algorithm, current reduction algorithms have common problem, such as results are single, algorithm execution time is long, search range is narrow. It is difficult to obtain minimum results and so on. So the paper put forward a novel attribute reduction algorithm--Based on Multi-Knowledge Mining PSO Reduction Algorithm.This algorithm combine the reduction algorithm and the particle swarm optimization algorithm to solve the search narrow range, single reduction results, algorithm execution time long, difficult to obtain minimum reduction problem. By using the UCI data sets to compare the new algorithm with the genetic reduction algorithm, the experimental results show the new algorithm has better performances in various aspects. In addition, for the rough set can't handle continuous attribute data, the paper realized the information entropy discrete data algorithm and applied it to the data preprocessing.Finally, the new algorithm is applied to factors analysis of violent crime. We make attempts to find the source causing violent crime happened for providing a new way for predicting violent crime.