Pre-Symptomatic Liver Disease Identification Using Distinct Machine Learning Methods

The liver which serves as a crucial human organ sustains metabolic operations while preserving health levels across the body. Successful treatment along with improved patient results require both quick and precise detection of liver disease through diagnosis. This study uses Machine Learning (ML) algorithms including Random Forest (RF), Categorical Boosting (CB), Adaptive Boosting (AB), Light Gradient Boosting Machine (LGBM) and Support Vector Classification (SVC) together with Logistic Regression (LR) to establish a method for Liver Disease (LD) prediction. Our research investigated numerous methods to assess their successful classification ability for liver disease with accuracy rate evaluations alongside precision and recall statistics and F1 score calculations. Notably, to make the reliability and generalizability of each model, the cross-validation was carried out. Our multi-algorithmic strategy boosts prediction stability by providing both precise analysis of each algorithm's predictive capabilities and their respective weaknesses in liver disease forecasting. CB due to its high accuracy and robustness, has produced the best results in comparison with other algorithms, which were analyzed. The ability to effectively deal with categorical variables and the low level of intensive preprocessing greatly contributed to its better performance in comparison with conventional models.