Predicting Covid-19 Trends Using Machine Learning based on Linear and Multiple Linear Regression Model

Abstract

The COVID-19 pandemic has adversely affected the health and economy of almost all the countries in the world including India. Almost thousands of people are getting affected by this daily. In this paper, analysis of the daily statistics of people who got affected and this proposed work is going to predict the future trend of the active cases in Odisha and India. Machine Learning based forecasting algorithms have proved their significance in generating predictive outcomes which are used to make decisions on actions that are going to happen in the future. ML algorithms have been using for a long time to do this kind of task. This proposed work is going to do analysis and prediction on the dataset which was created by COVID India organization. Linear and Multiple Linear Regression models are used to predict the future trend of active cases and also the number of active cases in fore coming days and to visualize the trend of future active cases. Here, the performance of Linear and Multiple Linear regression models are compared by using the R2 score. Linear and Multiple Linear regression got 0.99 and 1.0 as R2 scores respectively which shows that these are the strongest prediction models that are used to predict the future active cases of COVID - 19. Both these models acquired remarkable accuracy in COVID - 19 prediction. A strong correlation factor shows that there is a very strong relationship between a dependent variable (Active cases) and independent variables (positive, deceases, recovered cases)