Analysis of Trend in COVID-19, World Vaccination and Its Side Effects Using Machine Learning

Abstract

AbstractSARS-CoV-2 has created huge changes in the entire world. Trend of the outbreak of COVID-19 was constantly up during the initial days of the infection spread and gradually reduced at times. There were many trials on different varieties of vaccines manufactured by different companies. Presently precautions are taken to reduce the pandemic and significant results have been obtained for it and vaccines are also released for the security and welfare of mankind. The aim of the proposed work is to find the possible vaccine reactions, prediction of Covid-19 outbreak after the administration of vaccines using multiple machine learning algorithms, and visualization of few statistics regarding to the same. In this paper, we experimented with machine learning algorithms like support vectors, linear regression, and polynomial regression to predict vaccination. Classification algorithms like Support vectors with linear and RBF kernels, logistic regression, K-Neighbors Classifiers, Gaussian Naïve Bayes classifier, Decision Trees, and Random Forest algorithms for determining the vaccine reactions are implemented. We have achieved the highest accuracy of 0.56 for classifying the major symptoms after administering the vaccine using random forest classifier combined with optuna method of hyperparameter tuning and RMSE score of 0.091859 for number of people vaccinated using polynomial regression of degree four. To achieve this purpose, we have made use of Covid-19 disease data, World vaccination and Vaccine reactions dataset that were available on Kaggle.KeywordsSARS-CoV-2ClassifierRegressionHyperparameter tuning techniques