Classificação multirrótulo com aprendizado semissupervisionado: uma análise multivisão de dados

摘要

In the the last years, the computational techniques used for machine learninghave been divided or categorized according to the degree of supervision that exists inthese training’s set and according on the number of labels in this class attribute. Withinthese divisions, we find the semi-supervised learning, a technique that works well whennor all labels examples of the training set are known. In the other hand, the multi-labelclassification also is present in these categories and it proposes to classify exampleswhen they are associated with one or more labels. The combination of these learningtechniques generates the classification semi-supervised multi-label. Also in this context,there are sides that work with the semi-supervised learning for single vision and semisupervisedlearning data for multiple viewing data. The semi-supervised learningalgorithms for multiple viewing data has the basic idea of the exploitation ofdisagreements between the predictions of different classifiers, which is a subject rarelyaddressed in research. Thus, this work proposes the use of semi-supervised learning formulti-label classification using an approach with multiple viewing data, showing theresults of some experiments and comparing some results of experiments using the newmethods with the results of experiments using existing methods.