Heart Disease Risk Assessment by Using Light GBM Technique

Abstract

Heart disease is a term that covers various illnesses affecting the heart, which may involve blood vessels, heart rhythm, or congenital abnormalities. The World Health Organization (WHO) has identified heart disease as one of the primary causes of death globally. Heart disease includes conditions such as coronary artery disease and arrhythmias, among others. Predicting Heart disease risk can help doctors accurately assess patients' health, allowing for early intervention and lifestyle changes or medical treatment as needed. Machine learning (ML) can aid in understanding and reducing symptoms of heart disease by using key parameters such as heart rate, body temperature, and blood pressure. To more accurately predict heart disease risk and alert medical professionals and caregivers to patient location, the Light Gradient Boosting Machine (LightGBM) technique is proposed. LightGBM modelling is a promising classification strategy for predicting medication adherence in heart disease patients, assisting with patient stratification and decision-making based on the best available data. The chi-square statistical test is utilized to select specific features from the Cleveland heart disease (HD) dataset, and data visualization is used to depict feature relationships. In 303 instances of Cleveland HD dataset attributes, the random forest algorithm achieved an 88.5% accuracy rate during validation, according to experiment results.