Preemptive Diagnosis of Hepatitis C Using Machine Learning Techniques: A Retrospective Study in Saudi Arabia

Hepatitis C is an infectious chronic liver disease that can lead to liver cirrhosis and Hepatocellular Carcinoma (HCC) in its late stages. Those complications are associated with a high mortality rate and may necessitate liver transplants. Hepatitis C is a silent disease in its initial stages; therefore, a convenient tool for detecting the disease earlier before any symptoms appear is imperative to provide immediate treatment plans. Previous studies have used advanced laboratory tests, such as the Hepatitis C Ribonucleic Acid (RNA) test, to diagnose Hepatitis C, which are time-consuming and expensive. Therefore, this study utilized Saudi data obtained from King Fahad Specialist Hospital (KFSH) to preemptively diagnose Hepatitis C using demographic and simple laboratory tests. Machine Learning (ML) algorithms, including Decision Tree (DT), k-Nearest Neighbors (k-NN), and Naive Bayes (NB), were trained. To improve their results, their hyperparameters were tuned using GridSearchCV. Furthermore, the best feature subset for each model was found using the Genetic Algorithm (GA). Results have shown that k-NN with 11 features out of 44 outperformed other classifiers, with 86.05% accuracy, 92.3% recall, 85.7% precision, and 88.9% F1-score.