Detecting Heart Disease Using Machine Learning

Coronary diseases are a leading cause of global mortality, emphasizing the need for accurate and timely prediction methods. To reduce the death rate, the detection of heart disease at an early stage will be helpful. The United States witnesses a high incidence with one person being affected to heart disease every 34 seconds globally. In order to predict early-stage heart disease, machine learning model is built which identifies significant features through feature selection techniques. In the proposed work, UCI Cleveland dataset is used. The dataset comprises of 303 records with 13 features each. Three different approaches have been applied for feature selection such as Chi-Square, ANOVA F Value and Mutual Information. Then, three different machine learning algorithms like Logistic Regression, K Nearest Neighbors (KNN) and Support Vector Machine (SVM) are applied. Then, to optimize the performance of the algorithms, Ensemble method i.e. AdaBoost has been used. ML algorithms are evaluated using evaluation metrics such as accuracy, sensitivity, specificity, AUROC and log loss. In the proposed work, 91.80%, 91.80%, 91.80% and 88.52% are the accuracies achieved for SVM, KNN, AdaBoost and LR respectively. This study contributes to the ongoing efforts in developing accurate and interpretable models for early-stage detection of coronary diseases.