Predictive analysis on severity of Non-Alcoholic Fatty Liver Disease (NAFLD) using Machine Learning Algorithms

Fatty Liver Disease (FLD) is a frequent clinical impediment that is linked with high weariness and mortality. Despite that, an early prediction and diagnosis provide the patient with suitable treatment. For this, we aim to develop efficient Machine Learning (ML) models for the timely prognosis of FLD. We propose the use of Support Vector Machine (SVM), Logistic Regression, Random Forest (RF), Naive Bayes, and Multi-Layer Perceptron (MLP) on the dataset whose features are chosen by using the Mutual Information (MI) technique. This study uses the publically available dataset regarding FLD. This dataset is highly imbalanced, and to grapple with this, Synthetic Minority Oversampling Technique (SMOTE) was used. Results show that SVM performs well in comparison with the other state-of-the-art ML classifiers. In this study, we developed five models and compare the results with each other. Overall 99% accuracy is achieved by SVM and RF classification model.