The Epidemic Trend Of Typhoid And Evaluating The Predication Models In Zhenhai District, Ningbo City

Objectives:Study aims are to investigate the epidemic trend of typhoid in Zhenhai District,Ningbo city from 1985 to 2007 and to evaluate three prediction models-the ARIMA-time-series model,the neural net model and the GM(1,1) model.Methods:The study data of typhoid was derived from center of disease control and prevention,Zhenhai District,Ningbo city from 1985 to 2007.The description of epidemic trend of typhoid was preformed.Furthermore,the prediction models were generated from ARIMA-time-series model,the neural net model and the GM(1,1) model,using typhoid seasonal incidence from 1985 to 2004,respectively.And then three prediction models were assessed,using the typhoid seasonal incidence from 2005 to 2007.Results:The average of typhoid incidence was 23.42/100,000.The incidence was 24.68/100,000 and 22.08/100,000 among males and females,respectively.The data indicated the outbreak of typhoid happened in 1990(incidence=144.56/100,000).The fitness index of ARIMA-time-series model,the neural net model and the GM(1,1) model was 51.2%,80.6%and 11.96%,respectively.The sum of residual square of validating models was 5154.38,3559.24 and 646.58,using ARIMA-time-series model, the neural net model and the GM(1,1) model,respectively.Conclusion:This study indicated the incidence of typhoid was highest in Zhenhai District,Ningbo City.Hence, food administration has to be improved and typhoid vaccine should be widely used in this region.Among three prediction models,the neural net model was the best,and the ARIMA-time-series model was the next.