The coming of the post-genomic era has raised a growing demand for efficient procedures to identify protein functions, which can be accomplished by applying machine learning to the characteristics set extracted from the protein. This approach is called a feature-based approach and has been the focus of several works in Bioinformatics. In this work, we investigated the characteristics of proteins that improve the results found in the related works that used Support Vector Machine (SVM) in the enzymes class prediction. During the investigation, two approaches were evaluated: with and without dimensionality reduction, which was performed through the statistical technique Factor Analysis. Both approaches exceeded the results found in the related work, reaching F-measure averages of 85.10% and 83.90%, respectively, in the approachs with and without dimensionality reduction.