论文部分内容阅读
目的探讨自回归移动平均(autoregressive integrated moving average model,ARIMA)乘积季节模型在盐城市手足口病发病趋势预测的可行性。方法利用盐城市2009年1月至2015年12月的手足口病月发病率建立ARIMA乘积季节模型,并对2016年手足口病发病趋势进行预测。结果盐城市手足口病预测模型为ARIMA(1,0,1)(1,1,0)12,该模型的参数估计具有统计学意义,拟合优度检验统计量最小Normalized BIC=2.997,残差序列检验统计量Ljung-Box=20.692(P>0.05),残差为白噪声,模型能够拟合出手足口病的发病趋势,且实际值都在95%可信区间内,但模型拟合的平均误差率为41.296%,检验模型预测效果的平均误差率为23.998%,模型预测精度高于拟合精度。结论运用ARIMA乘积季节模型能够对盐城市手足口病发病趋势进行预测和动态分析,对手足口病预防控制产生积极的指导作用。
Objective To investigate the feasibility of forecasting the incidence of HFMD in Yancheng by using autoregressive integrated moving average model (ARIMA) product seasonal model. Methods The ARIMA product seasonal model was established based on the monthly incidence rate of HFMD in Yancheng from January 2009 to December 2015 and the prediction of the incidence of HFMD in 2016 was made. Results The prediction model of hand-foot-mouth disease in Yancheng was ARIMA (1,0,1) (1,1,0) 12. The parameter estimation of this model was statistically significant, with the smallest of the goodness-of-fit test statistic Normalized BIC = 2.997 The difference series test statistic Ljung-Box = 20.692 (P> 0.05), the residual is white noise, and the model can fit the trend of onset of hand-foot-and-mouth disease and the actual values are within 95% confidence interval, The average error rate is 41.296%, the average error rate of the test model prediction effect is 23.998%, and the prediction accuracy of the model is higher than the fitting accuracy. Conclusion ARIMA product seasonal model can predict and analyze the trend of HFMD in Yancheng and provide a positive guide for the prevention and control of HFMD.