Application Of Grass Roots Police Information Management Based On Decision Tree

The public security work is the leading force to maintain social order and crack down on the crime. How to effectively combat crime has been a concern of the society, and the police information management is an important basis to ensure the stability of social security. In the traditional mode of public security, according to the crime rate, security incidence rate, population of the police station to police the distribution. With the increase of population and the floating population, the original policing model can not effectively prevent crime, but also can not optimize the allocation of police resources. So new policing model arises of In order to prevent the replacement of combat, accurate crime prevention at the historic moment, crime prevention is the police station jurisdiction specific divided into different levels, for in different level of region, the different management methods. The goal is in the same police or shortage of police under the premise effectively combat crime, police resources optimization, reduce crime, improve social security situation. I-n recent years, many scholars and public security experts started in accordance with the various attributes of the social security situation of the same area is divided into different levels of security, namely public security management area classification.With the continuous development of computer technology, more and more new technologies and new algorithm is used in the security area in the management, domestic and international division of security region are mainly empirical method, statistics etc. There are some experts will be improved ant colony clustering algorithm, K means clustering algorithm and weighted fuzzy clustering algorithm, fuzzy c-means clustering algorithm, particle swarm optimization algorithm, and decision tree classification method of data mining technology is applied to the security area management. This paper studies the application of decision tree algorithm in the regional division of basic level public security management. The main work and research results of this paper are listed as follows:(1) Respectively, using k-means algorithm, DBSCAN algorithm, C4.5 algorithm of division of management areas. Through the experiment results prove that both K-means and DBSCAN algorithm, C4.5 algorithm for management division play a guiding role and C4.5 algorithm management areas data for discrete data processing, more accurate analysis of the main characteristics of the regional administration, to achieve the goal of precise division of regional.(2) The C4.5 algorithm on the basis of the research proposed simple C4.5 algorithm of rough set, rough set simple C4.5 algorithm is the improved algorithm of C4.5 algorithm and traditional C4.5 algorithm only rely on information gain rate selection decision points, classification based on a single, vulnerable to effects of category and the number of samples. Rough set expression under the environment of the data attributes of the decision effect of simple C4.5 algorithm, combining it with the decision maker of experience and knowledge can be more objective response attribute weights, to solve the C4.5 Classification Algorithm Based on a single, susceptible to classes in the sample set a number of shortcomings.(3) By using rough set contracted C4.5 algorithm to Dong feng county administrative data hierarchy of public security of Yellow River, considering the proportional distribution of the subjective and objective weights, to verify the feasibility of the algorithm in the division of management areas. Through the image analysis, we can see that the rough set C4.5 algorithm can better achieve the evaluation of regional division compared to the traditional C4.5, which shows that the rough set reduction C4.5 is feasible in the management area classification.(4) The rough set simple C4.5 is used in the data security of Yellow River Town, Dongfeng. The data(the annual amount of data into 21 groups, from criminal cases, public security case, distance, population number four attributes) in 2009 to 2015 for seven consecutive years of policing management of Yellow River Town, Dongfeng were predicted and analyzed by using C4.5 algorithm of decision tree, through visual graphic pattern contrast analysis results show that, after a comparative analysis, rough set C4.5 algorithm can achieve better prediction of policing management data analysis, with the actual local security situation in the region is relatively consistent.Studies have shown that using decision tree data mining algorithms are capable of implementation of community policing management data analysis, can science reasonable use of police resources, and realize the efficient means of crime prevention and precision strike crime, which will become the focus of research over the new policing model.