Machine Learning based Prediction of COVID-19: A Study on Italy's Pandemic Problems

Abstract

Starting from December 2019, COVID-19 has claimed several thousand lives all around the world. Though the novel Corona virus (SARS-CoV-2) has emerged from China, it has affected most of the western countries like European Union and America. Italy is one of the most affected countries accounting for 16523 deaths until 7th April 2020. This study involves machine learning based approach to analyse the effectiveness of lockdown in containing the COVID-19 problem. It is assumed that, the lockdown was implemented 9 previous dates and total affected cases were predicted accordingly. Also, total deaths were predicted for these conditions. From the results, it is evident that the rate of increase of total cases remained 1.128 on an average during 21 days lockdown starting from 9th March 2020. Furthermore, the rate decreased after the lockdown period. The model also predicted that a 9 days prior lockdown would have reduced the total deaths from 16523 to 2239. During a pandemics, like COVID-19 the essential action taken by any government should be a strict lockdown as early as possible. Also, the public should maintain proper social distancing to avoid community spread of a pandemic disease which is highly contagious like COVID-19.