An Actuarial Model Of Medical Insurance Fund Expenditure Based On Attribute Reduction Algorithm Of Roughset And Multivariate Linear Regression

Along with the Medical insurance system gradually mature and the rapid development of network technology, the data of the Medical insurance mass accumulation. In view of the medical insurance fund income and expenditure appeared a lot of models and methods. The research on health care policy adjustment and improve has played a very important role. But the domestic health care policy implementation of late, the research of medical insurance fund is not very perfect, the model is relatively simple, are more traditional methods. How to establish an accurate and efficient medical insurance actuarial model to remain a focus in the study of many medical insurance actuarial articles. The study of health care data analysis method also remains to be improved.First, An actuarial model of medical insurance fund expenditure according to Diseases is proposed in this paper. Previous actuarial model of medical insurance takes into account the worker identity, age, etc, caused by the pay scale is different. But it didn’t consider the distribution of various disease itself and the cost of time difference.To analysis and control the medical expense, must base on the analysis on medical expenses of single diseases. So this article will consider about the differences of various diseases in this model and propose An actuarial model of medical insurance fund expenditure according to Diseases. When calculating the process data, starting line and own parts is considered in this model, it make the model more realistic. At the same time single diseases expenses is also useful information for medical insurance fund expenditure analysis, so this model can avoid a lot of repetitive work.The analysis of the hospitalization cost is an important parameter in the model. The most widely in literature research method is a multivariate linear regression. But the choice of the independent variables has been the researchers according to their own research purposes, the lack of objectivity, and have higher request for researchers in related fields of knowledge. So this article tried to use the rough set attribute reduction algorithm to select the relevant influence factors. This method has more extensive applicability.Attributes reduction of rough set has also been a hotspot in the field of data mining. The RR (Random Reducts) algorithm is a frontier algorithm for large data, through the Random condition attribute column, the score calculated on the basis of the related algorithm to improve the efficiency of reduction. This paper makes improvement on the basis of the RR algorithm. Not only adds to the random of line but also will give up some score lower column in certain proportion every time after the calculation. To improve the accuracy and efficiency of the algorithm.The purpose of Theoretical research is practical application. Finally, the article Using this model to make predictions according to health care data of a city in 2004-2013. confirmed the feasibility of this method.