Many cellular functions are carried out in specific compartments of the cell. The prediction of the cellular localization of a protein is thus
related to its function identification. This paper uses two Machine Learning techniques, Support Vector Machines (SVMs) and Decision Trees,
in the prediction of the localization of proteins from three categories of organisms: gram-positive and gram-negative bacteria and fungi. For all
categories considered, the localization task has multiple classes, which correspond to the possible protein locations. Since SVMs are originally
designed for the solution of two-class problems, this paper also investigates and compares several strategies to extend this technique to perform
multiclass predictions.