Classification of NAFLD Using ML Algorithms

Nonalcoholic fatty liver disease (NAFLD) is a prevalent condition that can progress to permanent liver damage. Most current research focuses on identifying NAFLD patients with severe fibrosis (F3 and F4). This study aims to improve the predictive accuracy of three NAFLD stages: early (F0 and F1), intermediate (F2), and severe (F3 and F4), using machine learning (ML) techniques. Early detection of NAFLD can help prevent the disease from progressing to advanced stages or developing into liver cancer. Utilizing a dataset of 558 patients who underwent biopsies, we examined 6 classification algorithms and found that Random Forest outperformed other techniques, achieving 68.75% accuracy, F-1 score of 68.75%, weighted sensitivity of the model was 68.75%, positive predictive value was 68.85% and negative predictive value as 84.38%. Our method provides a more comprehensive prediction model for NAFLD stages, potentially leading to better outcomes in early detection and treatment