基于百度指数构建新冠感染境外输入预测模型的研究
Construction of an early warning model for imported cases of COVID-19 from aboard based on Baidu index
王舒颖 1于浩 2杨雪莹 2白羽 2曾强 2王欣2
作者信息
- 1. 300070,天津医科大学公共卫生学院
- 2. 天津市疾病预防控制中心
- 折叠
摘要
目的 通过分析新型冠状病毒感染相关搜索词与境外输入新型冠状病毒感染实际病例的相关性,考虑不同搜索词对境外输入新型冠状病毒感染传播的影响,构建符合我国具有境外输入特征的传染病模型.方法 选择2020年3月5日-2022年10月31日的百度相关搜索关键词和确诊新型冠状病毒感染境外输入病例的数据,采用相关分析,分析两者之间的相关性及时序变化特征,进而分别建立多元线性回归模型以及神经网络模型,并用均方误差(MSE)、均方根误差(RMSE)和拟合优度值(R2)评价两种模型的预测效果.结果 多元线性回归模型和神经网络模型均在提前3d预测时有较好的预测效果,且基于神经网络建立的预测模型拟合效果优于多元线性回归模型,提前3 d的MSE、RMSE以及R2的值分别为77.25、8.79和0.88.结论 根据百度指数关键词建立的神经网络模型对境外输入日新增新型冠状病毒感染病例有一定的预测能力,能够提前3 d预测病例的波动趋势,可作为境外输入新冠感染监测的补充手段.
Abstract
Objective To explore the correlation between search words related to coronavirus diseases 2019(COVID-19)and the acutal imported cases of COVID-19 from abroad with consideration of the impacts of different search words on the transmission of COVID-19 imported from abroad,so as to develop a model of imported infectious diseases suitable for China.Methods The related search keywords from March 5,2020 to October 31,2022 in Baidu and the data of confirmed overseas imported cases of COVID-19 were selected.The correlation analysis was used to analyze the correlation and temporal dynamics between the datasets to construct both a multivariate linear regression model and a neural network model.The predictive efficacy of the models were assessed by mean squared error(MSE),root mean squared error(RMSE),and the goodness-of-fit(R2).Results Both the multiple linear regression model and the neural network model showed good predictive performance when forecasting 3 days in advance.The prediction model based on the neural network exhibited better fitting effect than the multiple linear regression model and the MSE,RMSE,and R2 values of forcasting 3 d ahead were 77.25,8.79,and 0.88,respectively.Conclusions The neural network model established based on Baidu index keywords has a certain predictive ability for the daily new cases of COVID-19 imported from aboard and can predict the fluctuation trend of cases 3 days in advance,which can be used as a supplementary means for surveillance of imported COVID-19 from overseas.
关键词
百度指数/境外输入/新型冠状病毒感染/预测Key words
Baidu Index/Overseas import/COVID-19/Forecast引用本文复制引用
基金项目
天津市公共卫生科技重大专项(21ZXGWSY00010)
天津市医学重点学科(TJYXZDXK-066B)
天津市卫生健康科技项目(TJWJ2022MS046)
出版年
2024
NETL
NSTL
万方数据