Modelling and forecasting of new cases, deaths and recover cases of COVID-19 by using Vector Autoregressive model in Pakistan

Abstract

Highlights COVID-19 has increasing trend in Pakistan. Vector Autoregressive (VAR) is used for modelling and forecasting three key variables (new cases, deaths and recover cases) about COVID-19 in Pakistan. 10-days ahead forecast shows that Pakistan will face higher number of new cases and deaths. Recover cases is also higher.

Artificial neural networks for short-term forecasting of cases, deaths, and hospital beds occupancy in the COVID-19 pandemic at the Brazilian Amazon

“…Several studies have used ANN and other techniques to model a dynamic temporal of cases and deaths caused by COVID-19 in the world [ 48 – 50 ]. Saba & Elsheikh reported r yŷ like the cumulative variables of this study, after using nonlinear autoregressive ANN to model cumulative cases with data of 40 days [ 31 ].…”

Forecasting COVID-19 pandemic using optimal singular spectrum analysis

“…An ARIMA modelling has been utilized in [20] to forecast total infected cases of USA, Brazil, India, Russia, and Spain from 15th February to June 30, 2020. A Vector Autoregressive model has been used in [21] to forecast new daily confirmed cases, deaths and recovered cases in Pakistan for ten days. A Bayesian time series analysis has been conducted in [22] using daily data of COVID-19 in Japan until March 31, 2020.…”

Turn to the Internet First? Using Online Medical Behavioral Data to Forecast COVID-19 Epidemic Trend

“…In present study, we applied the VAR model for multivariate time series analysis to capture their dynamic interdependence by taking each variable as the linear function of past lags of itself and the past lags of the other explanatory variables ( Hamilton, 1994 ; Lütkepohl, 2006 ). Recent studies used VAR models to predict the trend of infection diseases such as sexually transmitted diseases ( Huang, Luo, Duan et al., 2020 ), Dengue ( Ramadona, Lazuardi, Hii et al., 2016 ) and the ongoing COVID-19 ( Khan, Saeed, Ali et al., 2020 ; Khan, Saeed, Ali et al., 2020 ; Fantazzini, 2020 ). A basic VAR model contains a set of n endogenous variables y t = ( y 1 t , y 2 t , …, y n t ).…”