灰色模型(1,1)和ARIMA模型在手足口病发病预测的应用  

Application of gray model(1, 1) and ARIMA model in predicting hand-foot-mouth disease

作　　者：原凌云[1,2] 于燕[1] 周以军[2] 焦欢[2] 张永英[2] 杨俊锋[2] 

机构地区：[1]西安交通大学医学部公共卫生学院,陕西西安710061 [2]安康市疾病预防控制中心,陕西安康725000

出　　处：《中国热带医学》2015年第12期1464-1467,共4页China Tropical Medicine

摘　　要：目的为了对安康市手足口病发病情况进行预测,优化拟合最佳传染病发病预测模型。方法采用2009—2014年手足口病的年发病率和月发病率分别建立灰色模型(1,1)和求和自回归滑动平均模型(ARIMA模型),并对两种模型发病预测结果进行比较、评价,优化拟合最佳预测模型,同时进行2015年安康市手足口病发病预测,并验证预测模型的准确度。结果 ARIMA模型对安康市手足口病预测效果优于灰色模型,2015年手足口病年发病率预测值为143.29/10万,每年发病分布存在两个发病高峰,分别在4—7月和11—12月。与2015年1—7月的实际发病率比较,ARIMA模型的发病率预测值略低,但无统计学差异(P>0.05)。结论采用ARIMA模型对手足口病发病预测效果优于GM(1,1)模型。Objective To predict morbidity of hand-foot-mouth disease in Ankang and optimize an optimal predictionmodel for infectious diseases. Methods Adopted annual morbidity and monthly morbidity of hand-foot-mouth disease from2009 to 2014 to establish gray model（1, 1） and Autoregressive Integrated Moving Average model（ARIMA model）, respectively,compared with prediction results of two models and evaluated results, optimized an optimal prediction model. Also we predictedhand-foot-mouth disease in Ankang in 2015 and verified accuracy of the prediction model. Results The prediction effects ofARIMA model in Ankang were superior to the gray model（GM）. The predicted value of hand-foot-mouth disease in 2015 was143.29/100 000. Annual morbidity distribution existed in two peaks, including April to July and November to December,respectively. Compared with actual morbidity from January to July in 2015, the predicted value of ARIMA model was slightlylower, but it had no statistical difference（P〉0.05）. Conclusion Prediction effects of ARIMA model on hand-foot-mouthdisease are better than GM（1,1）.

关 键 词：ARIMA模型 灰色模型 手足口病 预测 

分 类 号：R725.1[医药卫生—儿科]