Discovery On Association Rules And Sequential Patterns In Alarm Data Of Public Security

Data mining technology is one of the hot topics of current database and artificial intelligence research field, aimed at extracting valid, novel, potentially useful and easily understandable knowledge from the mass data. From the last century, the early 90's, data mining has been widely used in the market, financial investment, production and manufacturing, biotechnology, medicine, astronomy, geography and many other fields, but in public security work the research is still at the initial stage, has yet to mature technologies and applications. This paper focuses on applicating the data mining technology in the analysis of alarm data.In this paper, as the background of the online alarm and visiting alarm system, the system is running a large number of alarm data generated based on data mining technology research and development related applications. Through the analysis and comparations of relevant literature of data mining, we foucus on mining association rules and discovering sequential patterns in a large number of public security alarm data. Based on explaining the related concepts and issues models of association rules and sequential patterns, we focus on the design and implementation of the association rules algorithm and sequential patterns discovering algorithm. According to the practical application needs and the characteristics of the public security alarm data, we improve the algorithm of mining association rules and sequential patterns. In the pre-processing phase of the algorithm, we use the without guidance method to achieve the discretization of numerical attributes; multi-valued attributes will be splited equivalent; we use the partition strategies to reduce the amount of data. In the mining phase, respectively, we use the powered prefix and the weighted parameters to distinguish the importances of the attributes in order to dig out the rules of few incidents but serious case. Finally, we use simulation data to verify and analysis these algorithms. The association rules and sequential patterns which are minied by these algorithms have high value for the decision-making; deployment and forecasting of the police, at the same time, this information also provide an important thereunder of reference for the work of the emergency treatment.