Investigation of expert addition criteria for dynamically changing online ensemble classifiers with multiple adaptive mechanisms

This source preferred by Rashid Bakirov

Authors: Bakirov, R. and Gabrys, B.

Editors: Papadopoulos, H., Andreou, A., Iliadis, L. and Maglogiannis, I.

Pages: 646-656

Publisher: Springer Berlin Heidelberg

ISBN: 978-3-642-41141-0

DOI: 10.1007/978-3-642-41142-7_65

This data was imported from DBLP:

Authors: Bakirov, R. and Gabrys, B.

Editors: Papadopoulos, H., Andreou, A.S., Iliadis, L.S. and Maglogiannis, I.

https://doi.org/10.1007/978-3-642-41142-7

Journal: AIAI

Volume: 412

Pages: 646-656

Publisher: Springer

ISBN: 978-3-642-41141-0

This data was imported from Scopus:

Authors: Bakirov, R. and Gabrys, B.

Journal: IFIP Advances in Information and Communication Technology

Volume: 412

Pages: 646-656

ISBN: 9783642411410

ISSN: 1868-4238

DOI: 10.1007/978-3-642-41142-7_65

We consider online classification problem, where concepts may change over time. A prominent model for creation of dynamically changing online ensemble is used in Dynamic Weighted Majority (DWM) method. We analyse this model, and address its high sensitivity to misclassifications resulting in creation of unnecessary large ensembles, particularly while running on noisy data. We propose and evaluate various criteria for adding new experts to an ensemble.We test our algorithms on a comprehensive selection of synthetic data and establish that they lead to the significant reduction in the number of created experts and show slightly better accuracy rates than original models and non-ensemble adaptive models used for benchmarking. © IFIP International Federation for Information Processing 2013.

This data was imported from Web of Science (Lite):

Authors: Bakirov, R. and Gabrys, B.

Journal: ARTIFICIAL INTELLIGENCE APPLICATIONS AND INNOVATIONS, AIAI 2013

Volume: 412

Pages: 646-656

eISSN: 1868-422X

ISBN: 978-3-642-41142-7

ISSN: 1868-4238

The data on this page was last updated at 05:09 on February 27, 2020.