Authors:
Abeer A. Amer

Addresses:
Department of Computer and Information System, Faculty of Management Science, Sadat Academy for Management Science, Alexandria, Alexandria Governorate, Egypt.

Abstract:

This research applies machine learning (ML) techniques to predict liver disease using readily available blood test data, aiming to support early diagnosis and timely medical intervention. The study evaluates four widely used ML algorithms Random Forest, Support Vector Machine (SVM), Logistic Regression, and XG-Boost—using a dataset from the UCI Machine Learning Repository. To evaluate each model's effectiveness, several performance metrics were used, including the confusion matrix, precision, recall, F1 score, accuracy, and ROC_AUC, enabling a comprehensive comparison of classification performance. In the first phase of experimentation, the dataset was divided into 90% for training and 10% for testing. Under this configuration, the Random Forest model delivered the highest accuracy of 93%, followed by XG-Boost with 90%, while both Logistic Regression and SVM achieved 74% accuracy. Recognising the influence of data volume on model learning, the dataset was further split into 95% for training and 5% for testing. This modification resulted in notable performance improvements across all models. Random Forest achieved an impressive 98% accuracy, XG-Boost improved to 95%, Logistic Regression increased to 81%, and SVM rose to 79%, demonstrating the positive impact of additional training data on predictive performance.

Keywords: Chronic Liver Disease; Random Forest; Logistic Regression; Support Vector Machine (SVM); Machine Learning; Blood Test Data; Early Diagnosis; Confusion Matrix.

Received on: 18/06/2025, Revised on: 01/09/2025, Accepted on: 02/10/2025, Published on: 09/03/2026

DOI: 10.69888/FTSBE.2026.000672

FMDB Transactions on Sustainable Biomedical Engineering, 2026 Vol. 1 No. 1, Pages: 46–55

• Views : 61
• Downloads : 9

Download PDF