Statistical Analysis Of China's Total Cost Of Health Investment

In this paper, the analysis data was China's total health expenditure from 1978 to 2006, and two statistical models were used, namely, autoregression in-tegrated moving average model and autoregression with time varying coefficient model in the time series, to fit and analyze these data. In autoregression in-tegrated moving average model, we used PROC ARIMA procedure, the SAS statistical software package, to test, analyze and fit. On the promise of that model testing was applicable, we obtained a ARIMA (p, d, q) model for the health costs of our country, and according to the model, we coule predict the investment data of the health costs in the next decade. In autoregression with time varying coefficient model, we used the C language to obtain the estimates of the model. Simultaneously, according to the estimated value, we maked use of MATLAB software to generate 10 normal random numbers, and to combine the random numbers, we predicted the investment in China's health costs in the next decade. Finally, we overall merited and analyzed the prediction value of these two different statistical models, and the comparative methods was a simple sum of squares error method. Finally we obtained the rule of the investment of health costs of our country and drew the conclusion that auto regression with time varying coefficient model was better than autoregression integrated moving average model.