Machine Learning Approaches for Analysis of Covid-19 Data in India: A Case of Pandemic

Abstract

Covid-19 has made the whole world upside down with spreading of virus faster in various countries. India cases started in the month of March which panic all the peoples, yet the mortality rate (1.8%) is much less than the other countries. It is believed with native immunity of Indians that surveyed. But thou a dreathful time for the health care centre where the doctors and nurses spent sleepless night treating the cases. The lockdown has made relaxation in spread of the virus. Yet few states showed very high cases with the living culture and spread of virus were due to community spread too. In this study of Covid-19, machine learning techniques were applied to the datasets of twelve states with twelve dates assumed. The results were very promising with SVM, Naïve and DT models with accuracy of 100%. F1-score, precision and recall obtained as 1.0 whereas KNN accuracy was very poor with 60%. The confusion matrix accuracy obtained was 0.0821. CNN prediction is better over LSTM and Hybrid –LSTM – CNN models. Hence, the results proved that implementing ML and DL techniques would help to analysis the cases faster and monitor the region or states in the future any pandemic attacks.