Day-Level Forecasting of COVID-19 Transmission in India Using Variants of Supervised LSTM Models

Abstract

The novel Corona virus SARS-CoV-2 has started with strange new pneumonia of unknown cause in Wuhan city, Hubei province of China. On March 11, 2020, the World Health Organization declared the COVID-19 outbreak as a pandemic. Due to this pandemic situation, the countries all over the world suffered from economic and psychological stress. To analyze the growth of this pandemic, this paper proposes a supervised LSTM model and its variants to predict the infectious cases in India using a publicly available dataset from John Hopkins University. Experimentation has been carried out using various models and window hyper-parameters to predict the infectious rate ahead of a week, 2 weeks, 3 weeks and a month. The prediction results infer that, every individual in India has to be safe at home and to follow the regulations provided by ICMR and the Indian Government to control and prevent others from this complicated epidemic.