A pragmatic approach to multi-class classification

  • We present a novel hierarchical approach to multi-class classification which is generic in that it can be applied to different classification models (e.g., support vector machines, perceptrons), and makes no explicit assumptions about the probabilistic structure of the problem as it is usually done in multi-class classification. By adding a cascade of additional classifiers, each of which receives the previous classifier's output in addition to regular input data, the approach harnesses unused information that manifests itself in the form of, e.g., correlations between predicted classes. Using multilayer perceptrons as a classification model, we demonstrate the validity of this approach by testing it on a complex ten-class 3D gesture recognition task.

Export metadata

Additional Services

Share in Twitter Search Google Scholar
Metadaten
Author:Thomas Kopinski, Stéphane Magand, Alexander Gepperth, Uwe Handmann
URL:https://ieeexplore.ieee.org/document/7280768
URL:http://www.handmann.net/pdf/IJCNN-KopHanEtAl2015a.pdf
DOI:https://doi.org/10.1109/IJCNN.2015.7280768
ISBN:978-1-4799-1960-4
ISSN:2161-4407
Parent Title (English):The International Joint Conference on Neural Networks (IJCNN 2015)
Document Type:Conference Proceeding
Language:English
Year of Completion:2015
Release Date:2019/07/05
Volume:2015
Issue:2015
Pagenumber:8
First Page:1
Last Page:8
Institutes:Fachbereich 1 - Institut Informatik
DDC class:000 Allgemeines, Informatik, Informationswissenschaft / 004 Informatik
Licence (German):License LogoNo Creative Commons