A Web Application to predict the Liver Functionality using Ensemble Learning Technique

Abstract

Liver diseases pose a significant global public health challenge due to various causes like viral infections and metabolic issues. Early and precise diagnosis is crucial for effective management. While traditional diagnostic methods have limitations, the application of machine learning (ML) in medical diagnosis shows promise, enhancing predictive capabilities. By integrating Adaboost, Catboost, LGBM, and MLP classifiers into our approach, we aim to leverage their collective predictive power with a voting classifier. The goal is to improve the accuracy and reliability of liver disease prediction. This method achieved an 85% accuracy rate in determining whether a patient has a normal or pathological liver condition, generating two distinct outcomes. These findings suggest that the proposed system can complement a doctor’s diagnosis of liver disease.