Comparing Novel Multiple Logistic Regression to Bayesian Linear Regression for Accuracy in Death Ratio Analysis of Covid Patients

Abstract

Aim: Examining global data from different countries to estimate the covid infection fatality ratio. Materials and Methods: Accuracy of people who are suffering in different countries due to covid19 is analyzed. The sample size is 70 for both the algorithms and both these algorithms belong to supervised learning techniques. Each algorithm consists of 35 sample sizes from which calculate G-power which is 80%. All these algorithms come under supervised learning. Results: Calculated the total number of deaths and gained accuracy from the datasets and came to a conclusion that Novel Multiple Logistic Regression has an accuracy of 96% and Bayesian Linear Regression has an accuracy of 74%. The significance for accuracy is determined as p=0.042 (p<0.05). Conclusion:  Novel Multiple Logistic Regression has a better accuracy value than Bayesian LinearRegression.