ARIMA乘积季节模型在新疆肺结核发病预测中的应用

Application of ARIMA Multiplicative Seasonal Model in the Prediction of Pulmonary Tuberculosis Incidence in Xinjiang

摘要: 目的：根据新疆地区肺结核月发病数的季节性以及趋势性，建立求和自回归移动平均(ARIMA)乘积季节模型，并对新疆肺结核的发病趋势进行预测，调整防控措施。方法：利用R语言以2012年1月至2021年5月新疆地区肺结核每月发病人数为基础，建立并选出最适合的模型，对该地区2021年6月至2022年5月的肺结核发病人数进行一个预测，再将预测值与实际值作对比，以此为标准来讨论这个模型的预测效果。结果：通过赤池信息量(AIC = 46.23)与贝叶斯信息量(BIC = 57.1)最小原则可以得出，ARIMA(1, 1, 1)(1, 0, 0)₁₂是最优模型，2012年1月至2021年5月拟合结果，2021年6月至2022年5月模型预测值都落在置信区间95%内。结论：本文建立的ARIMA(1, 1, 1)(1, 0, 0)12能较为准确地预测新疆地区肺结核的月发病数。

Abstract: Objective: According to the seasonality and trend of the monthly incidence of pulmonary tuberculosis in Xinjiang, a Autoregressive Integrated Moving Average (ARIMA) multiplicative seasonal model was established to predict the incidence trend of tuberculosis in Xinjiang and adjust the prevention and control measures. Methods: Based on the monthly incidence of tuberculosis in Xinjiang from January 2012 to May 2021, the most suitable model is established and selected by using R language to predict the incidence of tuberculosis in this region from June 2021 to May 2022. Then, the predicted value is compared with the actual value, and the prediction effect of this model is discussed based on this standard. Results: ARIMA(1, 1, 1)(1, 0, 0)12 can be obtained by the principle of minimum Akachi information (AIC = 46.23) and Bayesian information (BIC = 57.1) is the optimal model. The fitting results from January 2012 to May 2021, and the predicted values of the model from June 2021 to May 2022 fall within the 95% confidence interval. Conclusions: ARIMA(1, 1, 1)(1, 0, 0)₁₂ established in this paper can accurately predict the monthly incidence of tuberculosis in Xinjiang.