Early Cardiovascular Risk Prediction in Quilombola Afro-descendants: A Data-Driven Approach

Abstract

Cardiovascular diseases (CVD) are the leading cause of global mortality. People from different social groups can be affected. However, socially vulnerable groups, such as the Quilombola communities in Brazil, may have an increased risk. Recently, sample data from this population were used to predict metabolic syndrome with machine learning(ML). Although metabolic syndrome is a risk factor for CVD, directly predicting cardiovascular risk itself might be more effective for implementing preventive strategies. Therefore, this study developed and assessed ML models to estimate CVD risk, including a variable importance analysis. Most models achieved over 80% effectiveness, with logistic regression achieving the best result. Considering the variable importance analysis, sex, age and income were identified as the most important variables, along with other socioeconomic and anthropometric data.

Ähnliche Arbeiten