Prediction of cardiovascular diseases with explainable artificialintelligence models

Artificial intelligence technologies have gained significant popularity in various fields, including engineering, defense, and education, as well as in the healthcare sector. Cardiovascular diseases are among the leading causes of death worldwide. Therefore, identifying risk factors for heart diseases and accurately predicting the risk of these diseases are of great importance. By analyzing datasets using artificial intelligence techniques, valuable insights can be obtained for disease diagnosis, treatment, and prediction. When these techniques are applied in high-risk domains such as healthcare, not only high accuracy but also explainability of the predictions is essential. This enables decision-makers to better understand the inner workings of the model and evaluate the accuracy and validity of the predictions. In the first few sentences, the importance of the subject and the aim of the thesis should be defined and brief information about the method and findings should be presented. Finally, the information produced should be expressed briefly. In this study, open-access datasets containing heart disease risk parameters, specifically "Heart Disease" dataset, were utilized to compare the performance of Logistic Regression, Random Forest, Extreme Gradient Boosting (XGBoost), Gradient Boosting (GB), and Light Gradient Boosting Machine (LightGBM) models using the Recursive Feature Elimination (RFE) method. Evaluation metrics included accuracy, precision, recall, and F1 score. As a result of the evaluation, the Logistic Regression algorithm achieved the highest accuracy rate of 91.67% after applying RFE. Additionally, the dataset's feature analysis was conducted using one of the explainable artificial intelligence methods, SHapley Additive exPlanations (SHAP). The analysis demonstrated the positive or negative impact of each feature on the target variable and presented their importance rankings.