Forecasting, Visualization and Analysis of COVID-19 in India using Time Series Modelling

EasyChair Preprint 4679

Abstract

Since the origination of COVID-19 in China and its spread across the globe, humanity has been put at risk and it has set a big alarm till its end across the country. Due to the unprecedented rate of increase in the number of cases and its subsequent pressure on the administration and health professionals globally, it would be highly needed to have a safe future by doing analysis and forecasting the number of new cases using some prediction methods. The current situation in India is getting worsened day-by-day due to which, the economy of this country has been down and unstable. In this paper, we have analyzed, how the numbers of daily infected cases in India could look like, predicting the trend, and investigate what the peak value could hit by now. We have used data-driven estimation methods like Fb-Prophet and long short-term memory (LSTM) as a state-of-the-art method and Deep Learning models respectively for forecasting the number of COVID-19 cases in India a few days ahead. We have proposed a method considering various parameters to predict daily confirmed future cases within a certain range which would be a beneficial tool for administrators and health officials.

Keyphrases: COVID-19, FB Prophet, Forecasting, LSTM, prediction, time series analysis

@booklet{EasyChair:4679,
  author    = {Afzal Ansari and Sourabh Kumar Burnwal},
  title     = {Forecasting, Visualization and Analysis of COVID-19 in India using Time Series Modelling},
  howpublished = {EasyChair Preprint 4679},
  year      = {EasyChair, 2020}}