Sequential Pattern Mining Applications In Medical Insurance

With the development of social health insurance, the medical information data of the insured expand rapidly. How to extract the valuable contents from the increasingly large medical databases and analyze the health insurance status become urgent requirements. This paper combines the sequential pattern mining algorithm with the health insurance database, and applies the cycle time constraint based sequential pattern mining algorithm in the health insurance database.At first, this paper analyzes the contents of current main sequential pattern mining algorithm, and for its characteristics this paper illustrates the limitations that the current popular Class of Apriori algorithm and the algorithm based on pattern-growth would do to the cycle of constraints based on time data. Then learn and to improve PCS_mine algorithm, and use the social health insurance medical database to verify that the algorithm has higher efficiency.The main research work has the following aspects.1. Deep studying and researching on preparation for data mining -Data preprocessing, understanding more about the mainstream multiple data sources and cleaning algorithm single data source integration. Then with NEU_cleaning Algorithm used by social health insurance in Zhengzhou City, this paper cleans the data from integrate multiple databases.2. In the analysis of constraint-based sequential pattern mining on the basis of various types, This paper focuses on time-based sequential pattern mining algorithm granularity PCS_mine and optimizes the algorithm to modify a HP-CSB data structure, Finally, it is verified by health insurance database, to show that the algorithm efficiency has been improved to some extent.