新疆医科大学学报

目的探讨ARIMA乘积季节模型在新疆地区猩红热月发病数预测中的应用,并对2018年新疆地区猩红热月新发病例数进行预测,为制订防控策略提供依据。方法对新疆地区2006-2016年猩红热月发病例数作为训练集,2017年猩红热月发病例数作为验证集,用训练集构建ARIMA乘积季节模型,随后用验证集检验模型精度。结果研究得到预测新疆地区猩红热月发病例数的最优模型为ARIMA(1,1,1)(0,1,1)₁₂,此模型训练集MAPE为34.81,验证集MAPE为20.52,此模型对新疆地区猩红热发病预测效果较为理想,预测出新疆地区2018全年一共可能有3 474例猩红热新发病人,其中月新发病数最大可能出现在11月,新发病例数可能为556例。结论通过构建ARIMA乘积季节模型对新疆地区的猩红热流行趋势进行预测是可行的,预测得到新疆地区2018年猩红热月新发病数处于较高水平。

Objective To explore the application of ARIMA product seasonal model in predicting the incidence of scarlet fever and the number of new cases of scarlet fever in Xinjiang. So as to provide evidence for formulating prevention and control strategies. Methods The number of scarlet fever cases in Xinjiang from 2006 to 2016 was used as the training set, and the number of scarlet fever cases in 2017 was used as the validation set. The ARIMA product seasonal model was constructed with the training set, and then the validation set was used to test the accuracy of the model. Results The optimal model for predicting the number of cases of monthly scarlet fever in Xinjiang was ARIMA(1,1,1)(0,1,1)₁₂. In this model, MAPE of training set and validation set were 34.81 and 20.52, respectively. The results showed that the model was effective in predicting the incidence of scarlet fever in Xinjiang. It was predicted that there may be 3 474 new cases of scarlet fever in Xinjiang in the whole year of 2018, of which the maximum number of new cases per month may occur in November and the number of new cases may be 556. Conclusion It is feasible to predict the epidemic trend of scarlet fever in Xinjiang by building ARIMA product seasonal model. It was predicted that the new incidence of scarlet fever in Xinjiang in 2018 is at a high level.