Dissertation > Excellent graduate degree dissertation topics show

The Research on Police Crime Analysis Using Data-mining Technology

Author: LiuHua
Tutor: ChenXiangTao；TanZuo
School: Hunan University
Course: Computer technology
Keywords: Data Mining Algorithm The weight parameter Sensitive parameters Priori parameters
CLC: TP311.13
Type: Master's thesis
Year: 2008
Downloads: 247
Quote: 2
Read: Download Dissertation

Abstract

The research and application of data mining techniques in public security work is still in its infancy, many public security business information system is still stuck at the level of primary treatment, the lack of comprehensive development and application of intelligent analysis and judgment, scientific decision-making warning; lack of data means by macro data is found to the problem of micro data from the micro to the macro processing capacity. How to use data mining techniques mining and extraction of hidden trends in the law have relevance in a large number of business data, improve the efficiency of public security law enforcement rapid response capability, in a timely manner to prevent and combat crime, to provide support services for the police decision, the focus of this study . The main work of this paper are as follows: the traditional Apriori algorithm in police crime analysis ignores the importance of different items, improved weighted association rule mining algorithm WARMA. Improved algorithm design model of weighted association rules, weighted by k-support expectations concept to exclude not become frequent k-itemsets candidate collection set, you can not solve a subset of the weighted association rule mining weighted frequent itemsets weighted frequent itemsets problem. The experiments and analysis show that: the algorithm WARMA has a set of candidate less and shorter execution time, and found significant crime more effective. Traditional Apriori algorithm can not be timely analysis of police crime keen to discover some new crime, based on the FUP improved incremental updating of association rules algorithm SFUP, SFUP algorithm by using a sensitive parameter to measure the attention of new projects to improve the general incremental updating algorithm for generate frequent itemsets process. The experiments and analysis show that: the improved algorithm is more obvious advantages the greater the amount of data that can be more effectively discovered a new type of criminal behavior, display space and time performance in the mining process. Traditional ID3 algorithm in the presence of police crime analysis, based on a priori parameters improved algorithm the BID3, BID3 algorithm makes the decision tree by increasing a priori parameters B, both? Shuo disaster plaque  Tiao edge seeking Yao? reducing the dependence of the value of more attributes, minimize large data mask small data occurs. The experiments and analysis show that: decision tree construction process to deal with the more massive data sets, BID3 algorithm ID3 algorithm more advantages than on efficiency and performance. On the basis of the actual combination of the police crime analysis, crime based on the decision tree algorithm design and experimental analysis of the prototype system, the functional requirements and system flow chart describes the prototype system modules and achieve. The experiments and analysis show that: mining processing huge amounts of data in the public security information database, use data mining techniques to discover trends law, auxiliary police decisions quickly and accurately, and in public security work has important practical significance.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer software > Program design,software engineering > Programming > Database theory and systems
© 2012 www.DissertationTopic.Net Mobile