Machine Learning-Based Approach to Predict Liver Disease Using an Imbalanced Dataset

Liver Disease describes a condition that damages the liver and can cause issues with its ability to operate. Liver disease affects an enormous percentage of people globally and it is currently one of the most common causes of death. It is important to diagnose liver disease as soon as possible in order to save human life and implement the necessary control measures. In this study, the emphasis lies on data pre-processing and feature extraction leading to improved performance for any machine learning classifier to predict liver disease. An array of Machine Learning models and ensemble approaches were assessed in the proposed system. Based on the outcomes from the study, it can be concluded that the Light Gradient Boosting Machine (LGBM) classifier perform better than the remaining models. It achieved 83.74% accuracy, 78% precision, 85% recall, 81% F1-score, and 90.3% AUC after utilizing the Synthetic Minority Oversampling Technique (SMOTE).