Research And Application Of Closed Itemsets Mining Algorithm In The Formulation Of Medical Insurance List

As an important part in the framework of China’s basic medical insurance system, medical insurance list provides the basic medical insurance coverage, and affects the protection level of the basic medical insurance. Therefore, formulating the medical insurance list reasonably is a key issue in the design of the basic medical insurance compensation mechanism. On the other hand, the gradual improvement of medical insurance information system makes a lot of healthcare data have been accumulated, thus promotes research and application of data mining techniques on healthcare data, and provides a new way to formulate the medical insurane list. Using data mining techniques to study the list formulation problem can take advantage of the large amount of valuable information from the healthcare data, which would make the medical insurance list more scientific and reasonable.Frequent closed itemsets mining is one of the most commonly used data mining techniques for discovering frequent patterns from massive datasets. It has been applied to many research areas and its algorithms have been extensively studied by scholars in both China and abroad. However, in the practical application, there are few parallel frequent closed itemsets mining algorithms that can easily add some constraints and have a high scalability to the explosively growing data. In this paper, on the basis of the algorithm CHARM which is used to mine frequent closed itemsets and can easily add constaints, we provide some theories to improve its shortcomings and then get algorithm NEWCHARM which is faster and more memory efficient. Besides, we also provide three methods to parallelize algorithm NEWCHARM so that we finally get algorithm PARACHARM, a parallel frequent closed itemsets mining algorithm that can easily add some constraints.This paper uses PARACHARM algorithm to study the problem of reasonably formulating the medical insurance list. Firstly, we define the problem of reasonably formulating the medical insurance list, and transform the real problem into a problem of mining frequent closed itemsets. Then, we extract and analyze the constraints of this problem, and design constraint-based PARACHARM algorithm to solve the problem. Finally, we conduct the experiment on the healthcare costs data after the data preprocessing. According to the experimental results, we extract a new medical insurance list, simulate the operation of this new list and analyze its running effect.The analysis indicates that the running effect of the new list is better than the random list, which can offer advice and reference for the formulation of medical insurance list.