Open AccessA comparative study of SIR Model, Linear Regression, Logistic Function and ARIMA Model for forecasting COVID-19 cases
Author(s) -
Saina Abolmaali,
Samira Shirzaei
Publication year - 2021
Publication title -
aims public health
Language(s) - English
Resource type - Journals
ISSN - 2327-8994
DOI - 10.3934/publichealth.2021048
Subject(s) - logistic regression , covid-19 , autoregressive integrated moving average , epidemic model , econometrics , statistics , linear regression , regression analysis , computer science , mathematics , medicine , time series , virology , outbreak , environmental health , population , disease , infectious disease (medical specialty)
Starting February 2020, COVID-19 was confirmed in 11,946 people worldwide, with a mortality rate of almost 2%. A significant number of epidemic diseases consisting of human Coronavirus display patterns. In this study, with the benefit of data analytic, we develop regression models and a Susceptible-Infected-Recovered (SIR) model for the contagion to compare the performance of models to predict the number of cases. First, we implement a good understanding of data and perform Exploratory Data Analysis (EDA). Then, we derive parameters of the model from the available data corresponding to the top 4 regions based on the history of infections and the most infected people as of the end of August 2020. Then models are compared, and we recommend further research.