Advanced search
1 file | 352.89 KB Add to list

On the role of cost-sensitive learning in multi-class brain-computer interfaces

(2010) BIOMEDIZINISCHE TECHNIK. 55(3). p.163-172
Author
Organization
Abstract
Brain-computer interfaces (BCIs) present an alternative way of communication for people with severe disabilities. One of the shortcomings in current BCI systems, recently put forward in the fourth BCI competition, is the asynchronous detection of motor imagery versus resting state. We investigated this extension to the three-class case, in which the resting state is considered virtually lying between two motor classes, resulting in a large penalty when one motor task is misclassified into the other motor class. We particularly focus on the behavior of different machine-learning techniques and on the role of multi-class cost-sensitive learning in such a context. To this end, four different kernel methods are empirically compared, namely pairwise multi-class support vector machines (SVMs), two cost-sensitive multi-class SVMs and kernel-based ordinal regression. The experimental results illustrate that ordinal regression performs better than the other three approaches when a cost-sensitive performance measure such as the mean-squared error is considered. By contrast, multi-class cost-sensitive learning enables us to control the number of large errors made between two motor tasks.
Keywords
cost-sensitive multi-class classification, BCI, kernel methods, ordinal regression, ORDINAL REGRESSION, SPATIAL FILTERS, ROC ANALYSIS, CLASSIFICATION, EEG

Downloads

  • (...).pdf
    • full text
    • |
    • UGent only
    • |
    • PDF
    • |
    • 352.89 KB

Citation

Please use this url to cite or link to this publication:

MLA
Devlaminck, Dieter et al. “On the Role of Cost-sensitive Learning in Multi-class Brain-computer Interfaces.” BIOMEDIZINISCHE TECHNIK 55.3 (2010): 163–172. Print.
APA
Devlaminck, Dieter, Waegeman, W., Wyns, B., Otte, G., & Santens, P. (2010). On the role of cost-sensitive learning in multi-class brain-computer interfaces. BIOMEDIZINISCHE TECHNIK, 55(3), 163–172. Presented at the NeuroMath Workshop Advanced Methods for the Estimation of Human Brain Activity and Connectivity.
Chicago author-date
Devlaminck, Dieter, Willem Waegeman, Bart Wyns, Georges Otte, and Patrick Santens. 2010. “On the Role of Cost-sensitive Learning in Multi-class Brain-computer Interfaces.” Biomedizinische Technik 55 (3): 163–172.
Chicago author-date (all authors)
Devlaminck, Dieter, Willem Waegeman, Bart Wyns, Georges Otte, and Patrick Santens. 2010. “On the Role of Cost-sensitive Learning in Multi-class Brain-computer Interfaces.” Biomedizinische Technik 55 (3): 163–172.
Vancouver
1.
Devlaminck D, Waegeman W, Wyns B, Otte G, Santens P. On the role of cost-sensitive learning in multi-class brain-computer interfaces. BIOMEDIZINISCHE TECHNIK. 2010;55(3):163–72.
IEEE
[1]
D. Devlaminck, W. Waegeman, B. Wyns, G. Otte, and P. Santens, “On the role of cost-sensitive learning in multi-class brain-computer interfaces,” BIOMEDIZINISCHE TECHNIK, vol. 55, no. 3, pp. 163–172, 2010.
@article{981658,
  abstract     = {Brain-computer interfaces (BCIs) present an alternative way of communication for people with severe disabilities. One of the shortcomings in current BCI systems, recently put forward in the fourth BCI competition, is the asynchronous detection of motor imagery versus resting state. We investigated this extension to the three-class case, in which the resting state is considered virtually lying between two motor classes, resulting in a large penalty when one motor task is misclassified into the other motor class. We particularly focus on the behavior of different machine-learning techniques and on the role of multi-class cost-sensitive learning in such a context. To this end, four different kernel methods are empirically compared, namely pairwise multi-class support vector machines (SVMs), two cost-sensitive multi-class SVMs and kernel-based ordinal regression. The experimental results illustrate that ordinal regression performs better than the other three approaches when a cost-sensitive performance measure such as the mean-squared error is considered. By contrast, multi-class cost-sensitive learning enables us to control the number of large errors made between two motor tasks.},
  author       = {Devlaminck, Dieter and Waegeman, Willem and Wyns, Bart and Otte, Georges and Santens, Patrick},
  issn         = {0013-5585},
  journal      = {BIOMEDIZINISCHE TECHNIK},
  keywords     = {cost-sensitive multi-class classification,BCI,kernel methods,ordinal regression,ORDINAL REGRESSION,SPATIAL FILTERS,ROC ANALYSIS,CLASSIFICATION,EEG},
  language     = {eng},
  location     = {Leuven, Belgium},
  number       = {3},
  pages        = {163--172},
  title        = {On the role of cost-sensitive learning in multi-class brain-computer interfaces},
  url          = {http://dx.doi.org/10.1515/BMT.2010.015},
  volume       = {55},
  year         = {2010},
}

Altmetric
View in Altmetric
Web of Science
Times cited: