用于预测新冠病毒确诊病例的“早期R”流行模型和时间序列模型分析
Analysis of 'earlyR' epidemic model and Time Series model for prediction of COVID-19 registered cases.
作者信息
Kanagarathinam Karthick, Algehyne Ebrahem A, Sekar Kavaskar
机构信息
Department of EEE, GMR Institute of Technology, Rajam, Andhra Pradesh, India.
Department of Mathematics, Faculty of Sciences, University of Tabuk, Saudi Arabia.
出版信息
Mater Today Proc. 2020 Oct 14. doi: 10.1016/j.matpr.2020.10.086.
The COVID-19 is an epidemic that causes respiratory infection. The forecasted data will help the policy makers to take precautionary measures and to control the epidemic spread. The two models were adopted for forecasting the daily newly registered cases of COVID-19 namely 'earlyR' epidemic model and ARIMA model. In earlyR epidemic model, the reported values of serial interval of COVID-19 with gamma distribution have been used to estimate the value of R and 'projections' package is used to obtain epidemic trajectories by fitting the existing COVID-19 India data, serial interval distribution, and obtained R0 value of respective states. The ARIMA model is developed by using the 'auto.arima' function to evaluate the values of (p, d, q) and 'forecast' package is used to predict the new infected cases. The methodology evaluation shows that ARIMA model gives the better accuracy compared to earlyR epidemic model.
新冠病毒病是一种引起呼吸道感染的流行病。预测数据将有助于政策制定者采取预防措施并控制疫情传播。采用了两种模型来预测新冠病毒病的每日新增病例，即“earlyR”疫情模型和自回归积分移动平均（ARIMA）模型。在“earlyR”疫情模型中，已使用具有伽马分布的新冠病毒病序列间隔报告值来估计R值，并使用“projections”软件包通过拟合现有的印度新冠病毒病数据、序列间隔分布以及各邦获得的R0值来获得疫情轨迹。ARIMA模型是通过使用“auto.arima”函数来评估（p, d, q）值而开发的，并使用“forecast”软件包来预测新的感染病例。方法学评估表明，与“earlyR”疫情模型相比，ARIMA模型具有更高的准确性。
Zotero 插件