指数平滑模型在上海市猩红热发病率预测中的应用

孔德川; 潘浩; 郑雅旭; 姜晨彦; 韩若冰; 吴寰宇; 陈健

doi:10.3784/j.issn.1003-9961.2019.10.016

指数平滑模型在上海市猩红热发病率预测中的应用

Application of exponential smoothing model in predicting incidence of scarlet fever in Shanghai

摘要

摘要:
目的探讨指数平滑模型在上海市猩红热月报告发病率预测的应用。
方法根据2004 — 2017年上海市猩红热月报告发病率的时间分布，选择Holt-Winters指数平滑模型拟合2004 — 2017年6月上海市猩红热报告发病率资料，并利用最优模型对2017年7 — 12月的猩红热报告发病率进行预测。
结果猩红热报告月发病率经自然对数变换后的Holt-Winters加法指数平滑模型的拟合相对最优，决定系数（R²）＝0.884，标准化贝叶斯信息准则值为−2.568，其残差为白噪声序列，预测值与实际值基本吻合，2017年7 — 10月的预测精确度优于11—12月。
结论指数平滑模型可用于上海市猩红热月报告发病率的短期预测，4个月内的短期预测效果更优。

Abstract:
ObjectiveTo explore the application of exponential smoothing model in the prediction of monthly incidence of scarlet fever in Shanghai.
MethodsAccording to the time distribution of monthly reported incidence of scarlet fever in Shanghai from 2004 to 2017, Holt-Winters exponential smoothing model was used to fit the incidence data of scarlet fever in Shanghai from 2004 to June 2017, and the optimal model was used to predict the incidence of scarlet fever in Shanghai from July to December 2017.
ResultsThe fitting effect of Holt-Winters additive exponential smoothing model of scarlet fever monthly reported incidence after natural logarithmic transformation was relatively optimal, the R² was 0.884, and the standardized bayesian information criterion value was −2.568. Its residual was a white noise sequence. The predicted value was basically consistent with the actual value, and the predicted accuracy for the period from July to October was better than that for the period from November to December in 2017.
ConclusionThe exponential smoothing model can be used for the short term prediction of monthly incidence of scarlet fever in Shanghai,and the short-term prediction effect in four months is better.

HTML全文

参考文献(19)

施引文献

资源附件(0)