부산대학교 도서관

Globally, the leading causes of death are cardiovascular diseases (CVD). Although clinicians have begun to pay more attention to one of the CVD indications, heart failure, current clinical practice typically falls short of obtaining high accuracy in such occupations. Machine learning has advantages for feature rating as well as clinical prediction, which improves how clinical professionals comprehend the outcomes. Therefore, the concept of artificial intelligence aims to find the issue of machine learning models lack of answerable in the medical domain and provide medical professionals with patient-trained decision-making tools that enhance therapies and diagnostics. This study builds a model for predicting heart problems survival using tree based algorithm and machine learning. Extreme Gradient Boosting (XGBoost) is one of the tree-based algorithms demonstrated to give the most accurate results (83% accuracy with unknown data). A features selection preprocessing is also carried out to verify which related features contribute to the results of the model.