《厦门大学学报（自然科学版）》

研究新型冠状病毒肺炎(COVID-19)传播方式规律及趋势有助于有效遏制其蔓延.介绍了一些常用的传染病预测模型,提出了Logistic_SEIR模型,既克服了Logistic模型不能预测现有确诊人数的缺点又克服了SEIR模型调参太多的缺点,并通过实验证明了所提出模型的实现和预测的优越性.同时进一步研究了Logistic_SEIR模型需要调试的参数取不同的初始值对预测结果的影响,并通过加权误差来量化分析预测效果.最后指出未来进一步的研究方向.

To study the law and trend of coronavirus disease 2019(COVID-19)transmission mode is helpful for effectively curbing its spread.In this article,we first introduce some commonly used prediction models of infectious diseases,and then introduce the Logistic_SEIR model.This model not only overcomes shortcomings of the Logistic model that cannot predict the number of confirmed patients,but also overcomes those of the SEIR model in which plenty of parameters exist.In experiments,we prove the superiority of the realization and prediction of the model proposed in our study.At the same time,we further study the influence of different initial values of the parameters that need to be debugged in the Logistic_SEIR model on the prediction results,and quantifies the prediction effect through the weighted error.Finally,in the summary,we point out the future research directions.

引言
1 传染病预测常用方法
2 Logistic与SEIR结合模型
3 实验和结果分析
4 结论

图1 SEIR模型示意图<br/>Fig.1 Schematic diagram of SEIR model

图1 SEIR模型示意图
Fig.1 Schematic diagram of SEIR model

图2 Logistic_SEIR模型预测现有确诊人数<br/>Fig.2 Logistic_SEIR model prediction of the number of existing infected population

图2 Logistic_SEIR模型预测现有确诊人数
Fig.2 Logistic_SEIR model prediction of the number of existing infected population

表1 SEIR模型与Logistic_SEIR模型的RMSE<br/>Tab.1 RMSE of SEIR model and Logistic_SEIR model

表1 SEIR模型与Logistic_SEIR模型的RMSE
Tab.1 RMSE of SEIR model and Logistic_SEIR model

图3 不同初始健康人数S0的预测曲线<br/>Fig.3 Prediction curves of different initial health population numbers S0

图3 不同初始健康人数S0的预测曲线
Fig.3 Prediction curves of different initial health population numbers S0

表2 不同S0初始值下误差计算<br/>Tab.2 Error calculation of different initial values of S0

表2 不同S0初始值下误差计算
Tab.2 Error calculation of different initial values of S0

图4 不同转化率α的预测曲线<br/>Fig.4 Prediction curves of different conversion rates α

图4 不同转化率α的预测曲线
Fig.4 Prediction curves of different conversion rates α

表3 不同α初始值下误差计算<br/>Tab.3 Error calculation of different initial values of α

表3 不同α初始值下误差计算
Tab.3 Error calculation of different initial values of α

图5 不同康复率γ的预测曲线<br/>Fig.5 Prediction curves of different rehabilitation rates γ

图5 不同康复率γ的预测曲线
Fig.5 Prediction curves of different rehabilitation rates γ

表4 不同γ初始值下误差计算<br/>Tab.4 Error calculation of different initial values of γ

表4 不同γ初始值下误差计算
Tab.4 Error calculation of different initial values of γ

图6 额外测试集上不同康复率γ的预测曲线<br/>Fig.6 Prediction curve of different rehabilitation rates γ on additional test sets

图6 额外测试集上不同康复率γ的预测曲线
Fig.6 Prediction curve of different rehabilitation rates γ on additional test sets

[6] LIPSITCH M,COH EN T,COOPER B,et al.Transmi-ssion dynamicsand control of severe acute respiratory syndrom[J].Science,2003,300(562):1966-1970.
[7] RILEY S,FRASER C,DONNELLY C A,et al.Transmission dynamics of the etiological agen t of SARS in HongKong:impact of public health interventions[J].Science,2003,300(5627):1961-1965.
[8] DYE C,GAY N.Modeling the SARS epidemic[J].Science,2003,300(5627):1884-1885.
[9] GONG J H,SUN Z L,LI X W,et al.Simulation and analysis of SARS control[J].Journal of Remote Sensing,2003,7(4):260-265.
[10] 龚建华,周洁萍,徐珊,等.SARS传播动力学模型及其多智能体模拟研究[J].遥感学报,2006,10(6):832-834.
[11] 赵旭,陈立萍,程维虎.Logistic回归模型在人口问题中的应用[J].应用概率统计,2015,31(6):662-666.
[12] 原存德,胡宝安.具有阶段结构的SI传染病模型[J].应用数学学报,2002(2):193-203.
[13] 韩翔,夏苏徽.基于SIS模型的禽流感传染模型[J].现代商贸工业,2018,39(22):80-81.
[14] 王香阁,于红斌,王飘飘,等.基于SIR模型对埃博拉病毒蔓延趋势的预测[J].福建电脑,2015,31(2):97,115.
[15] 喻孜,张贵清,刘庆珍,等.基于时变参数-SIR模型的2019-nCoV疫情评估和预测[J].电子科技大学学报,2020,49(3):357-361.
[16] 曹盛力,冯沛华,时朋朋.修正SEIR传染病动力学模型应用于湖北省2019冠状病毒病(COVID-19)疫情预测和评估[J].浙江大学学报(医学版),2019,49(2):178-184.
[17] 范如国,王奕博,罗明,等.基于SEIR的新型肺炎传播模型及拐点预测分析[J].电子科技大学学报,2020,49(3):369-374.

备注

引言

1 传染病预测常用方法

2 Logistic与SEIR结合模型

3 实验和结果分析

4 结论

学报简介

备注

引言

1 传染病预测常用方法

2 Logistic与SEIR结合模型

3 实验和结果分析

4 结 论

学报简介

4 结论