Application of an autoregressive integrated moving average model for predicting injury mortality in Xiamen, China

摘要: Objective Injury is currently an increasing public health problem in China. Reducing the loss due to injuries has become a main priority of policies. Early warning injury mortality based on surveillance information essential for reducing or controlling disease burden injuries. We conducted this study find possibility applying autoregressive integrated moving average (ARIMA) models predict from Xiamen. Method The monthly data Xiamen (1 January 2002 31 December 2013) were used fit ARIMA model with conditional least-squares method. values p, q and d (p, d, q) refer numbers lags, lags differences, respectively. Ljung–Box test was measure ‘white noise’ residuals. mean absolute percentage error (MAPE) between observed fitted evaluate predicted accuracy constructed models. Results A total 8274 injury-related deaths identified during period; annual rate 40.99/100 000 persons. Three models, (0, 1, 1), (4, 0) (1, (2)), passed parameter (p 0.05) tests, MAPE 11.91%, 11.96% 11.90%, chose 1) as optimum model, value which similar that other but fewest parameters. According there would be 54 persons dying each month 2014. Conclusion could applied