Time Series Analysis of COVID-19 Data- A study from Northern India

Abstract

The continuing new Coronavirus (COVID-19) pandemic has caused millions of infections and thousands of fatalities globally. Identification of potential infection cases and the rate of virus propagation is crucial for early healthcare service planning to prevent fatalities. The research community is faced with the analytical and difficult real-world task of accurately predicting the spread of COVID-19. We obtained COVID-19 temporal data from District Surveillance Officer IDSP, Dehradun cum District Nodal Officer- Covid-19 under CMO, Department of Medical Health and Family Welfare, Government of Uttarakhand State, India, for the period, March 17, 2020, to May 6, 2022, and applied single exponential method forecasting model to estimate the COVID-19 outbreak's future course. The root relative squared error, root mean square error, mean absolute percentage error, and mean absolute error were used to assess the model's effectiveness. According to our prediction, 5438 people are subjected to hospitalization by September 2022, assuming that COVID cases will increase in the future and take on a lethal variety, as was the case with the second wave. The outcomes of the forecasting can be utilized by the government to devise strategies to stop the virus's spread.

Comparison of exponential smoothing and ARIMA time series models for forecasting COVID-19 cases: a secondary data analysis

“…( 20 Dealing with medical data is a tedious task, especially when working on forecasting COVID-19 data as the virus and its variants have shown a dramatically change over the period. 21 Different models have been introduced to forecast pandemic outbreak since it's arise. A study compared various time series models to ten different countries while estimating the proportion of active cases in the overall population.…”