This paper shows the preliminary results of a simulation study devoted to comparing, in a multi-class classification setting, three classifiers that transform the probabilities produced by a probabilistic classifier into a single class: the usual Bayes Classifier and the new Max Difference Classifier and Max Ratio Classifier. As well known, the Bayes Classifier has some limits with rare classes, whereas the proposed Max Difference and Max Ratio Classifiers seem to represent better alternatives.
CATEGORICAL CLASSIFIERS IN MULTI-CLASS CLASSIFICATION PROBLEMS
Maurizio Carpita;Silvia Golia
2021-01-01
Abstract
This paper shows the preliminary results of a simulation study devoted to comparing, in a multi-class classification setting, three classifiers that transform the probabilities produced by a probabilistic classifier into a single class: the usual Bayes Classifier and the new Max Difference Classifier and Max Ratio Classifier. As well known, the Bayes Classifier has some limits with rare classes, whereas the proposed Max Difference and Max Ratio Classifiers seem to represent better alternatives.File in questo prodotto:
Non ci sono file associati a questo prodotto.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
