Web-Based Heart Disease Prediction by Comparison and Implementation of SVM, AdaBoost, and Hybrid SVM-AdaBoost Algorithms

Heart disease is a global health problem and the main cause of death worldwide. In Indonesia, the prevalence of heart disease is still on the rise, accounting for 6.1% of the total population. It is crucial to concentrate on early detection and prevention in order to lessen the effects of this condition. Machine learning algorithms have been used in earlier studies to forecast the likelihood of getting heart disease. This study aims to compare various classification approaches, including a single classification using the Support Vector Machine (SVM) algorithm, the AdaBoost algorithm, and a hybrid classification that combines the SVM and AdaBoost algorithms. UCI Cardiac Disease data will be used by adopting the CRISP-DM methodology. The best model used to develop website prototypes using the Flask framework. The findings show that the SVM-AdaBoost hybrid algorithm outperforms the single classification algorithm in terms of accuracy. The single SVM algorithm and single AdaBoost algorithm show an accuracy rate of 86.67%, while the SVM-AdaBoost hybrid algorithm shows an accuracy rate of up to 90.00%. Therefore, the use of SVM and AdaBoost hybrid classification algorithms is recommended to predict the likelihood of developing heart disease.