Early Detection of Heart Disease Based on Medical Check-Up Datasets Using Multilayer Perceptron Classifier

Abstract Background Heart disease is a leading cause of mortality worldwide, including in Indonesia. With millions of deaths attributed to heart disease annually, early detection and prevention are essential in reducing its impact. This study aims to address the issue of late diagnosis by focusing on early detection and monitoring of heart disease symptoms. Risk factors such as high blood pressure, high cholesterol, smoking, obesity, and sedentary lifestyle contribute to the development of heart disease. Findings Many patients with heart disease are diagnosed at advanced stages, indicating a lack of timely detection. To overcome this, the study utilizes Medical Check-Up Datasets and applies the Multilayer Perceptron (MLP) Classifier with back-propagation for early diagnosis. The MLP Classifier employs supervised learning during training and testing, where a value of 1 represents abnormal (presence of heart disease) and 0 represents normal (absence of heart disease). The study achieves an accuracy of 99.2% in early detection. Conclusion The findings of this study demonstrate the feasibility and significance of early detection in heart disease diagnosis. By using the MLP Classifier, healthcare providers can promptly identify and address heart disease symptoms, enabling appropriate treatment planning and reducing the risk of complications. These results provide valuable insights for healthcare professionals, contributing to improved patient care and outcomes.