Using Co-training and Self-training in Semi-supervised Multiple Classifier Systems

摘要: Multiple classifier systems have been originally proposed for supervised classification tasks, and few works dealt with semi-supervised multiple classifiers. However, there are important pattern recognition applications, such as multi-sensor remote sensing multi-modal biometrics, which demand able to exploit both labelled unlabelled data. In this paper, the use, in systems, of two well known learning methods, namely, co-training self-training, is investigated by experiments. Reported results on benchmarking data sets show that self-training allow exploiting different types classifiers systems.