Analysis of Influencing Risk Factors for Covid-19 Infection Based on the Predictive Models Using Machine Learning Algorithms

Abstract

Over recent years, machine learning has turned very reliable in the medical field. Prediction of COVID-19 by using Machine Learning algorithms would help to increase the speed of disease identification resulting in reduced mortality rate.This work focuses on therole of machine learning algorithms for the identification of risk factors that influences the prediction of COVID-19 infection. This will enable the people to be cautious about the risk factors and directs for proper medical treatment. COVID-19 data provided by WHO with 5434 patient’s records and 19 features collected from kaggle.com is considered for this study.Collected data is preprocessed to make it suitable for further analysis. Dependent features are identified and channelized based on the influencing risk factor through exploratory data analysis. Based on the influencing factors, Predictive models are built by considering 70% of data for training and the remaining 30% for testing. These models are evaluated and experimental results are analyzed thatindicates that the Random Forest algorithm gives higher accuracy by minimizing RMSE. This work also endorses that the higher accuracy 93.25% with minimum RMSE value 25.96% is attained when the prediction is based on the symptom. Hence this work concludes that symptom is the higher influencing risk factor that needs to be addressed promptly during COVID-19 infection.