Detection of Nonalcoholic Fatty Liver Disease Using Deep Learning Algorithms

Abstract

Some occasional drinkers develop Nonalcoholic Fatty Liver Disease (NAFLD). Hepatocytes are the key indication of NAFLD. Western nations are seeing rising non-alcoholic fatty liver disease (NAFLD). About 25% of Americans have this chronic liver condition. Recent research estimates that 33.66 percent of Bangladeshi adults have fatty liver disease, affecting over 45 million people. This illness is a major cause of liver-related deaths. Thus, minimizing fatty liver disease risk is crucial. Failure to diagnose fatty liver early may cause serious medical consequences. This study examines fatty liver signs and disorders to help diagnose diabetes early. This study shows the association between fatty liver symptoms and illness to help diagnose early. Deep learning categorization methods are widely utilized to build patient risk prediction models. In this study, “used” was utilized. This article uses numerous deep learning approaches to predict fatty liver disease. Convolutional, Long Short-Team Memory, Recurrent, and Multilayer perception neural network designs were mentioned. This study calculates AUC, shows correlation matrices, and visualizes features, and the optimum method. Deep learning achieved 71% accuracy in a highly categorized environment.