时间序列模型在广州市登革热发病趋势预测的应用

Application of time series model in predicting the incidence of dengue fever prevalence trend in Guangzhou

目的利用ARIMA模型对广州市登革热发病趋势进行时间序列分析和预测。方法收集广州市2015-2018年登革热每周发病数,采用2015-2017年登革热每周发病数构建ARIMA模型,通过2018年登革热发病预测值与实际值的拟合情况,评价模型预测效果。结果拟合模型ARIMA(4,1,9),残差序列为白噪声,预测值与实际值平均绝对误差为4.03,均方根误差为8.13。2018年登革热预测发病趋势与实际发病趋势较吻合。结论 ARIMA模型能较好地模拟广州市登革热的短期发病趋势,可作为预测工具。

Objective To analyze and predict incidence of dengue fever(DF)in Guangzhou with an auto regressive integrated moving average(ARIMA)model.Methods DF weekly incidence was collected from 2015 to 2018,and a ARIMA model was fit using data from 2015 to 2017.The prediction performance was evaluated by comparison of reported and predicted DF incidences in 2018.Results The model ARIMA(4,1,9)was established with residual sequence of a white noise sequence.The average absolute error between the predicted and actual values was 4.03,and the root mean square error was 8.13.The predicted epidemic trend in 2018 was consistent with the reported data.Conclusions The established ARIMA model can predict the shot-time trend of DF in Guangzhou with good performance.It can be used as an applicable tool for prediction.