Diagnosis of Heart Disease Using Binary Grasshopper Optimization Algorithm and K-Nearest Neighbors

The heart is one of the main organs of the human body, and its unhealthiness is an important factor in human mortality. Heart disease may be asymptomatic, but medical tests can predict and diagnose it. Diagnosis of heart disease requires extensive experience of specialist physicians. The aim of this study is to help physicians diagnose heart disease based on hybrid Binary Grasshopper Optimization (BGO) Algorithm and K-Nearest Neighbors (KNN). The BGO algorithm is used for feature selection (FS), and the KNN is used for classification.

In this study, the medical records of 270 patients in the field of heart disease with 13 features were evaluated. The number of patients is equal to 120 and the absence of disease is equal to 150, so the data set is balanced. Patient information is taken from the standard UCI (University of California, Irvine) database. The evaluation of the proposed model has been done in MATLAB simulation.

According to the evaluations, the accuracy was 89.82%, the sensitivity was 89.61%, and the specificity was 90.41%, which are acceptable compared to the results of previous studies in the field of heart disease. Also, the percentage of accuracy of the proposed method based on 7 features (Age, Sex, Chest Pain, BP, Electrocardiographic, Angina, and Thallium) is equal to 90.35%.

According to the results of this study, for the diagnosis of heart disease, the proposed method has been more effective in diagnosing the disease and selecting important features in comparison with previous methods.