Prediction of Early Stage of Fatty Liver Disease in Patients using Logistic Regression and Naive Bayes Algorithm

Abstract

The main objective of the work is to evaluate the accuracy in predicting fatty liver disease at an earlier stage in patients by using the Novel Logistic Regression (LR) algorithm and Naive Bayes (NB) algorithm. The dataset is taken from CHAOS Grand challenge website. It consists of a liver image dataset. The sample size per group is calculated as 16. The pretest power is calculated as 95% alpha value set as 0.05. The Novel Logistic Regression classifier is used to identify fatty liver disease at an early stage in patients and compared with Naive Bayes classifier in terms of accuracy. The Naive Bayes classifier produces 76.9% accuracy in predicting the fatty liver disease at an early stage whereas the Novel Logistic Regression classifier accuracy is 83.3%. The significant value is 0.994. Hence Novel Logistic Regression is better than Naive Bayes. The performance of Novel Logistic Regression is better compared to Naive Bayes in terms of both accuracy and precision.