The Design And Implementation Of Automatic Categorization System Of Public Security Information Based On SVM

This thesis comes from the system of "Public Security Information System", which is corporately developed by the police office of Zhengzhou, Hennan province, and electron business affairs center of Henan province. This system covers the range of Zhengzhou city and the county under the domination of Zhengzhou. It sets up a share and process system for the police office of Zhengzhou. Through the developing and debugging, this system has already run in the reality and operated in order. This paper aims at the real requirement of text categorization of public security information and the text characteristic of public security information, and deeply studies the technology of text categorization and the theory of Support Vector Machine (SVM). Sequentially, Automatic Categorization System of Public Security Information based on SVM is realized and designed, the capability of this system is optimized. The practice shows that this system achieves the good effect of categorization.The main research content and the innovations of the paper is the following:1. On the basis of deeply study the SVM, aimed at SVM parameter selection problems,this paper gives parameters optimization method based on the genetic algorithm, This method simplifies the process of SVM parameter multi-attempts, relizes SVM parameter optimization automatic.2. According to the characteristic of the text of public security information, this paper introduces the regional weight modulus in the process of Text Vector Representation, improves the classic TF-IDF. It makes the weight of lemma can full-scale express the contributions on distincting the lemma's category.3. This paper proposes the system model according with the object of public security information. The subcomponents' relation is loose coupling in the model. On one hand, this system facilitates selecting the middle result of the categorization process and supporting for public security information's analysis; on the other hand, this system reduces the coupling degree of the subcomponents in the system, provides the flexible system structure, and improves the adaptability of system.4. This paper designs and realizes the Automatic Categorization System of Public Security Information based SVM. The rationality of the design for this system is verified by practice. At last, the key issue of the implement process of the system is discussed.