The approach based on an ensemble of SVMs offered a satisfactory performance, improving the results when compared to a single SVM model using the same features, and showed better results in comparison with previous machine learning approaches of the state-of-the-art.