Abstract:
BACKGROUND & OBJECTIVE: Statistical models play a significant role in understanding the trend, level and trajectory of infectious diseases and provide the foundation to formulate efective policies and timely intervention, so that the morbidity and mortality due to these diseases can be declined. This study aimed to uncover the trend and proposing a forecasting model for daily expected outbreaks due to COVID-19 of fourth spike in Pakistan. METHODOLOGY: This study is primarily based on a secondary data of COVID-19 daily confirmed outbreaks. The twomonth (1st June to 31st July 2021) time series data is recorded and available from COVID-19 health advisory platform by Ministry of National Health Services Regulation and Coordination oficial website. Descriptive and time series analysis (ARIMA, exponential smoothing models) were applied. The analysis was carried out using R programming language. RESULTS: The highest (5026) and the lowest (663), COVID-19 confirm cases reported on 31 July 2021 and 21 June 2021 respective, whereas the average confirmed cases were 1830 [762-2898] per day. Four diferent time series models are executed namely ARIMA, Brown, Holt and Winter. Among competitive models, ARIMA (0, 2, 1) is found to be an optimum forecasting model, selected by using auto ARIMA function with least root mean square error. A day ahead forecast is obtained under the selected ARIMA model and yielded that COVID-19 confirmed outbreaks is expected to increase about 3.1% per day. CONCLUSION: COVID-19 outbreaks are expected to rise in Pakistan and ARIMA (0, 2, 1) is an optimum forecasting model for daily COVID-19 outbreaks.
Keywords:
Pakistan
,
Forecast
,
COVID19
,
ARIMA
,
Time series analysis