151. Learning from Imbalanced Datasets with Cross-View Cooperation-Based Ensemble Methods
- Author
-
Sokol Koço, Cécile Capponi, Laboratoire d'Informatique et Systèmes (LIS), Aix Marseille Université (AMU)-Université de Toulon (UTLN)-Centre National de la Recherche Scientifique (CNRS), éQuipe d'AppRentissage de MArseille (QARMA), Aix Marseille Université (AMU)-Université de Toulon (UTLN)-Centre National de la Recherche Scientifique (CNRS)-Aix Marseille Université (AMU)-Université de Toulon (UTLN)-Centre National de la Recherche Scientifique (CNRS), Springer, and ANR-15-CE23-0026,LIVES,Apprendre avec des vues interactionnelles(2015)
- Subjects
Boosting (machine learning) ,Computer science ,business.industry ,boosting ,ensemble methods ,Novelty ,Multi-class Learning ,Confusion matrix ,Imbalanced Learning ,02 engineering and technology ,Machine learning ,computer.software_genre ,Ensemble learning ,imbalanced classes ,[INFO.INFO-AI]Computer Science [cs]/Artificial Intelligence [cs.AI] ,confusion matrix norm ,020204 information systems ,0202 electrical engineering, electronic engineering, information engineering ,sort ,020201 artificial intelligence & image processing ,Multi-view Learning ,Artificial intelligence ,business ,computer ,Classifier (UML) - Abstract
International audience; In this paper, we address the problem of learning from im-balanced multi-class datasets in a supervised setting when multiple descriptions of the data-also called views-are available. Each view incorporates various information on the examples, and in particular, depending on the task at hand, each view might be better at recognizing only a subset of the classes. Establishing a sort-of cooperation between the views is needed for all the classes to be equally recognized-a crucial problem particularly for imbalanced datasets. The novelty of our work consists in capitalizing on the complementariness of the views so that each class can be processed by the most appropriate view(s); thus improving the per-class performances of the final classifier. The main contribution of this paper are two ensemble learning methods based on recent theoretical works on the use of the confusion matrix's norm as an error measure, while empirical results show the benefits of the proposed approaches.
- Published
- 2019