Comparative Study of Gradient Boosting Algorithms for Heart Disease Prediction
Supervisor Name
Emad Natsheh
Supervisor Email
e.natsheh@najah.edu
University
An-Najah National University
Research field
Artificial Intelligence
Bio
Dr. Emad Natsheh is an Associate Professor in the Faculty of Engineering and Information Technology at An-Najah National University. He holds a PhD in Artificial Intelligence and Renewable Energy and an MSc in Computer and Network Technology from Manchester Metropolitan University (UK), and a BSc in Computer Engineering from An-Najah National University. His research focuses on artificial intelligence, machine learning, smart systems, and renewable energy applications. Dr. Natsheh has published several papers in international journals and conferences and has experience in training in data science, Python programming, and data analysis.
Description
Machine learning offers methods for analysing clinical data and predicting the presence of heart disease. However, model performance can depend on both data preprocessing and hyperparameter selection. This study will compare four gradient boosting algorithms—XGBoost, LightGBM, CatBoost, and HistGradientBoosting—and examine the separate and combined effects of one-hot encoding and hyperparameter optimisation using Optuna. Using a heart disease dataset containing patient information, each algorithm will be evaluated under four experimental configurations: default hyperparameters without external one-hot encoding, tuned hyperparameters without external one-hot encoding, default hyperparameters with external one-hot encoding, and tuned hyperparameters with external one-hot encoding. The categorical feature handling used in each configuration will be documented to support interpretation of the comparisons. Performance will be assessed using accuracy, precision, recall, specificity, F1-score, and ROC-AUC. All configurations will use consistent evaluation splits, with preprocessing and hyperparameter tuning restricted to training partitions within a nested stratified cross-validation procedure. The analysis will assess whether performance differences depend on the algorithm, encoding strategy, or their interaction with hyperparameter tuning. The study aims to identify the best-performing configuration on the selected dataset and provide evidence to guide the selection of gradient boosting models and preprocessing strategies for heart disease classification.