Heart disease is one of the important diseases that causes many human casualties in different countries of the world every year. However, early diagnosis of heart disease is very important, because the patient can be treated faster. In Iraq, working on medical diagnoses is a necessity and a challenge. To detect heart disease, angiography method is expensive and has significant side effects. But artificial intelligence (AI) and machine learning (ML) models can help meet this need. However, AI algorithms and ML models may suffer from low accuracy. To deal with the existing challenges, in this research, a heart disease detection algorithm based on the hierarchical method of 4 machine learning models is presented. The hierarchical method makes the accuracy of the proposed method increase and the results of more models are used in decision making. Therefore, in the proposed method by combining the results of disease diagnosis in ML, SVM (Support Vector Machine), ANN (Artificial Neural Network), DT (Decision Tree), KNN (K-Nearest Neighbors) models, and using the selection algorithm The SFS (Sequential Forward Selection) feature provides an accurate and reliable system. Also, in the proposed method, the use of SFS has led to the selection of important features and increased accuracy. Simulation results on the UCI database show that the proposed method detects disease with an accuracy of 92.5%.
