Performance Analysis of Support Vector Machine and Gradient Boosting Machine Algorithms for Heart Disease Prediction

Abstract

Cardiovascular disease ranks among the primary causes of mortality globally, with death rates rising each year. Assessing heart disease risk is crucial for enhancing the efficiency of prevention and treatment strategies. This study seeks to evaluate the effectiveness of two machine learning techniques, namely Support Vector Machine and Gradient Boosting Machine, in forecasting heart disease using a dataset obtained from Kaggle. The research process starts with gathering data, followed by exploratory analysis, preprocessing through label encoding, handling class imbalance with SMOTE, and normalizing data using Standard Scaler. Features were selected using the Correlation Thresholding method. Subsequently, the dataset was divided into training and testing sets to develop predictive models. The model performance was assessed using evaluation metrics, including accuracy, precision, recall, and F1-Score. The findings indicate that the Gradient Boosting Machine outperformed the Support Vector Machine, achieving an accuracy of 98% compared to SVM's accuracy of 93%. This research is expected to contribute to healthcare practices by enabling early detection of heart disease risks. Future research is recommended to explore other algorithms or employ more diverse datasets to achieve better results