Previsão da quantidade de classes em classificação hierárquica multirrótulo

Abstract

Many Machine’s Learning classification problems are described in the literature associanting the classification of data with a belonging class to an finite set of classes, all at a same level. However, many classification problems are of hierarchical nature, in which classes may be subclasses or superclasses of other classes. In many hierarchical problems, one or more examples may be associated to more than one class simultaneously. Those problems are known as multi-label hierarchical classification problems. In this paper, the ML-kNN techniques used to address the prediction of multi-label problems, aiming to determine the number of classes that may be assigned to an example.