Bareilly College, Bareilly, Uttar Pradesh, India 243005
bSchool of Computational and Intergative Sciences, Jawaharlal Nehru University, New Delhi, India 110067
cCSIR-Central Electronics Engineering Research Institute,Pilani Rajasthan, India 33303

Abstract

The COVID-19 pandemic, combined with uncertainties about the efficacy of current vaccines and medical procedures used in the fight against coronavirus, has disrupted all aspects of life worldwide, necessitating the urgent need for robust forecasts to assist humanity in bringing this global and unprecedented crisis under control. Using Google Trends data, this study employs three long short-term memory (LSTM) models: Stacked LSTM, Bi-LSTM, and Vanilla LSTM, to predict the test positivity rate of COVID-19 in the United States. The findings of the current study were promising, as the three models could closely capture the trend of the positive rate of infections over the study duration, which was consistent with the findings of other studies.