A Comparative Analysis of Decision Tree Classifier Performance in the Medical Data Analysis

Accurate classification of diseases, such as heart disease and breast cancer, is crucial in the medical field. It enables early detection, appropriate treatment planning, and improved patient outcomes, emphasizing the need for precise classification models. This study aims to evaluate and compare the performance of three widely used decision tree classifiers, namely ID3, C4.5, and CART, using the Statlog heart disease dataset and the Wisconsin Breast Cancer Database (WBCD). Additionally, it explores the impact of oversampling, feature selection, and hyperparameter tuning on the Machine Learning(ML) pipeline. Results indicate that C4.5 consistently outperforms ID3 and CART in both datasets. Furthermore, integrating oversampling, feature selection, and hyperparameter tuning techniques with the C4.5 classifier notably improves F1 scores on the datasets. These findings underscore the significance of these components in enhancing classification efficiency, highlighting strategies for robust ML pipelines tailored to medical data.