Predicción de riesgo cardiovascular a través de los algoritmos supervisados: Support Vector Machine, Gradient Boosting y Naïve Bayes

Abstract

Abstract: The primary objective of this titration work (TT) is to predict cardiovascular risk using supervised Machine Learning algorithms. Three models were evaluated, Support Vector Machine (SVM), Gradient Boosting (GB), and Naïve Bayes (NB), analyzing their performance through evaluation metrics such as accuracy, recall and F1-Score, in the four experiments carried out, a normalization technique was applied using StandardScaler, and class balancing techniques such as SMOTE and Random Over Sampling to improve the minority classes. The GB model, in experiment number four, was the most effective model, reaching an accuracy of 0.87%, which worked with normalized and balanced data, SVM was the second model with the best results in its evaluation metrics, while the NB model worked with unnormalized, but balanced data, where it presented more difficulties compared to the other models, presenting very low metrics. It is important to carry out an adequate analysis of the data and select the most optimal model with the best metrics, which will allow us to identify cardiovascular risk more accurately and quickly.