2011 | OriginalPaper | Buchkapitel
A Novel Classifier Ensemble Method Based on Class Weightening in Huge Dataset
verfasst von : Hamid Parvin, Behrouz Minaei, Hosein Alizadeh, Akram Beigi
Erschienen in: Advances in Neural Networks – ISNN 2011
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
While there are many methods in classifier ensemble, there is not any method which uses weighting in class level. Random Forest which uses decision trees for problem solving is the base of our proposed ensemble. In this work, we propose a weightening based classifier ensemble method in class level. The proposed method is like Random Forest method in employing decision tree and neural networks as classifiers, and differs from Random Forest in employing a weight vector per classifier. For evaluating the proposed weighting method, both ensemble of decision tree and neural networks classifiers are applied in experimental results. Main presumption of this method is that the reliability of the predictions of each classifier differs among classes. The proposed ensemble methods were tested on a huge Persian data set of handwritten digits and have improvements in comparison with competitors.