CARDIOVASCULAR RISK ASSESSMENT USING CLASSIFICATION MODELS

Abstract

According to the World Health Organization (WHO): every year, approximately 37 million people in the world suffer a cardiovascular event and approximately 46% of those people die from these causes. It is known that the occurrence of a cardiovascular event results from the interaction of different risk factors in which we can include hypertension, high levels of blood lipids, diabetes, obesity and smoking. The diagnosis of cardiovascular risk can be made from the presence of any of these risk factors; moreover, there are pocket calculators that allow estimating cardiovascular risk from a model proposed by Framingham, which is essentially a generalized linear model. In the present work, in addition to using the Framingham model, other classification models are used such as Decision Trees, logistic regression and Random Forest. The objective is to choose the best classification model based on goodness criteria such as the correct classifications rate, relative efficiency and deviance.