Please use this identifier to cite or link to this item: http://hdl.handle.net/2381/37337
Title: Dealing with Multiple Classes in Online Class Imbalance Learning
Authors: Wang, S.
Minku, Leandro Lei
Yao, X.
First Published: 12-Jul-2016
Presented at: 25th International Joint Conference on Artificial Intelligence (IJCAI'16), 9th-15th July, 2016, New York City, USA
Publisher: International Joint Conferences on Artificial Intelligence
Citation: 25th International Joint Conference on Artificial Intelligence (IJCAI'16) 2016
Abstract: Online class imbalance learning deals with data streams having very skewed class distributions in a timely fashion. Although a few methods have been proposed to handle such problems, most of them focus on two-class cases. Multi-class imbalance imposes additional challenges in learning. This paper studies the combined challenges posed by multi-class imbalance and online learning, and aims at a more effective and adaptive solution. First, we introduce two resampling-based ensemble methods, called MOOB and MUOB, which can process multi-class data directly and strictly online with an adaptive sampling rate. Then, we look into the impact of multi-minority and multi-majority cases on MOOB and MUOB in comparison to other methods under stationary and dynamic scenarios. Both multi-minority and multi-majority make a negative impact. MOOB shows the best and most stable G-mean in most stationary and dynamic cases.
Links: http://www.ijcai.org/Abstract/16/302
http://hdl.handle.net/2381/37337
Embargo on file until: 1-Jan-10000
Version: Post-print
Status: Peer-reviewed
Type: Conference Paper
Rights: Copyright © 2016, International Joint Conferences on Artificial Intelligence. The file associated with this record is under a permanent embargo while permission to archive is sought from the publisher. The full text may be available through the publisher links above.
Appears in Collections:Conference Papers & Presentations, Dept. of Computer Science

Files in This Item:
File Description SizeFormat 
ijcai16-final-submitted.pdfPost-review (final submitted)538.14 kBAdobe PDFView/Open


Items in LRA are protected by copyright, with all rights reserved, unless otherwise indicated.