Heart Boost: Clinical Data-Driven Heart Disease Prediction Using XGBoost

Abstract

Cardiovascular diseases, especially heart disease, remains one of the main reasons for increasing mortality rate globally. Timely diagnosis plays a critical role in preventing complications and thus lessening the load on healthcare systems. In this research, we presented an interpretable and accurate machine learning framework for estimating the likelihood of heart disease using the Cleveland Heart Disease dataset from the UCI Machine Learning Repository. The dataset comprises extensive patient-related clinical features, such as age, sex, chest pain type, resting blood pressure, cholesterol levels, fasting blood sugar, electrocardiographic results, maximum heart rate achieved, exercise-induced angina, ST depression, and thalassemia. This dataset was first cleaned and pre-processed to handle missing values and encrypt categorical variables. Thereafter, we used three machine learning algorithms—Logistic Regression, Random Forest, and Extreme Gradient Boosting (XGBoost)—were skilled and examined. We carried out model evaluation utilizing established performance metrics which includes accuracy, precision, recall, F1-score, and the area under the ROC curve (AUC). Among the models examined, XGBoost topped the others with a prediction accuracy of 93%, demonstrating its strong generalization and robustness. XGBoost not only achieved highest accuracy but also performed well in all the other evaluation metrices such as F1-score, precision, recall. Through this study we analyzed that machine learning, when applied appropriately with interpretable techniques, can successfully assist heart disease diagnosis. This work emphasizes the potential of integrating clinical data with ML models to develop reliable, transparent, and scalable diagnostic tools that can help physicians in early detection and medication design.