Research Of Time Series Predictive Model On Incidence Of Three Kinds Of Respiratory Infectious Disease

ObjectiveAim at the incidence of three kinds of respiratory infectious disease and their factors ,the more reasonable predictive model will be investigated And established ,which will offer decision-making foundation to prevention and surveillance , at the same time which will offer scientific reference to establish prediction model of other respiratory infections.MethodsApplying time series analysis( including :ARIMA multiplicative model, exponential smoothing method and seasonal periods model ), grey forecast gm(1,1) model to analyze the incidence of three kinds of respiratory infectious disease (including: Tuberculosis,Epidemic parotitis,Measles) from 2003 to 2007 year, and to establish their predictive models . Finally, the established models will be estimated,evaluated ,then the best predictive model will be established.ResultThe incidence of disease (including: Tuberculosis,Epidemic parotitis Measles) is seasonally fluctuant-distributing and has tendency variation from 2003 to 2007 year. The epidemic parotitis broke out on December 2006 and January 2007 , which leaded to the incidence of the epidemic parotitis higher than other months . The Measles had a massive epidemic occurrence in 2007 year,which caused a high incidence in this year.The results of the ARIMA model and exponential smoothing method is good. Comparing the exponential smoothing method with ARIMA model ,the ARIMA model's results is better than exponential smoothing method, so the ARIMA model may reasonably predict in future. The gray GM(1,1) model and the seasonal periods model are good in analysis of seasonal incidence of three kinds of respiratory infectious disease, but which are the best reasonable is not able to be judged.