A Novel Coronavirus Prediction Model Based on Naïve Bayes

Abstract

Any area of life, including medicine, is quickly impacted by technological breakthroughs. The use of AI to analyse data and make decisions has shown encouraging outcomes in the healthcare industry. COVID-19 has quickly spread to over 100 nations. The future repercussions might affect people everywhere. Creating a control system that can identify the coronavirus is of the utmost importance. Disease diagnostics using a variety of AI techniques might be one way to regulate the present chaos. In this study, we used classical and ensemble ML algorithms to categorise textual clinical reports into four groups. Word bag, report length, and term frequency/inverse document frequency (TF/IDF) were some of the methods used for feature engineering. Both ensemble and standard machine learning classifiers were fed these characteristics. With a testing accuracy of 96.2%,logistic regression and multinomial naive bayes outperformed other ML methods. A recurrent neural network may provide more precise results in the future.