Toward transductive learning classifiers for non-stationary EEG

Raza, H; Prasad, G; Li, Y; Cecotti, H

Abstract

A major challenge in brain-computer interfaces (BCIs) based on electroencephalogram (EEG) signals is the immanent non-stationarities in EEG data. Statistical properties of the signals may shift from inter-or-intra session transfer that often led to deteriorated BCI performance. We propose to handle the issue with a transductive learning approach. The performance of the proposed method is evaluated on BCI competition 2008-Graz dataset B. The results show an improvement in classification accuracy over the traditional learning method.

Citation

Raza, H., Prasad, G., Li, Y., & Cecotti, H. (2014, August). Toward transductive learning classifiers for non-stationary EEG. Presented at 36th Annual International Conference of the Institute of Electrical and Electronics Engineers (IEEE) Engineering in Medicine and Biology Society, Chicago, Illinois, USA

• Presentation Conference Type: Other
• Conference Name: 36th Annual International Conference of the Institute of Electrical and Electronics Engineers (IEEE) Engineering in Medicine and Biology Society
• Conference Location: Chicago, Illinois, USA
• Start Date: Aug 26, 2014
• End Date: Aug 30, 2014
• Publication Date: Aug 28, 2014
• Deposit Date: Jun 19, 2015
• Publisher: Institute of Electrical and Electronics Engineers
• Additional Information: Event Type : Conference