基于人口流动和时空信息的城市疫情影响研究

An Investigation of COVID-19 Based on Population Flow and Spatio-temporal Information

新型冠状病毒(COVID-19)自暴发以来对人类的身体健康以及国家乃至整个世界的经济产生巨大影响.因此,如何研究及预测COVID-19的传播对于疾病防控尤为重要.本文基于人口迁移和城市间的距离,采用机器学习(即线性回归和随机森林模型)对COVID-19肺炎的确诊人数进行预测.在进行预测之前,我们对相关特征与疫情数据进行了相关性分析.结果表明,疫情数据与人口迁移和距离两种因素之间均表现出较强的相关性.综合两个模型来看,随着目标城市迁徙指数高的城市和邻近城市个数增加,预测结果的精度会不断提高.在机器学习预测中,随着作为特征的目标城市迁徙指数高的城市个数的增加,线性回归模型的预测性能变差,但随机森林模型的性能变好.并且,随机森林的预测效果一直优于线性回归.综上所述,人口迁徙和距离等相关数据有助于提高新型冠状病毒(COVID-19)肺炎的确诊人数预测精准度,为新型冠状病毒(COVID-19)肺炎或者其他传染性疾病确诊人数的预测提供了一种新的思路.

The outbreak of COVID-19 has largely affected people's daily life and the economy of our country as well as the whole world.Therefore,it is essential to study and predict the transmission of COVID-19 for the purpose of disease prevention and control.Basing on the population migration and city distance,this paper studies and predicts COVID-19 by introducing machine learning algorithms,i.e.,linear regression and random forest.Before making the prediction,it analyzes the correlation characteristics with the epidemic data.The results show that the epidemic data has a strong correlation between both population migration and distance factors.Combining with the two models,with the increasing number of high migration index cities and neighboring cities,the accuracy of the prediction results will be continuously improved.In the machine learning prediction,with the increasing number of high migration index cities,the prediction performance of the linear regression model deteriorates,but the performance of the random forest model becomes better.Moreover,the prediction effect of random forests is always better than linear regression.In summary,the data related to population migration and distance can help to improve the accuracy of the prediction of COVID-19 and provide a new idea for predicting the number of confirmed cases of COVID-19 or other contagious disease.