Abstract
A new formulation of principal component analysis (PCA) that considers group structure in the data is proposed as a Variable Subset Selection (VSS) method. Optimization of electrode channels is a key problem in brain-computer interfaces (BCI). BCI experiments generate large feature spaces compared to the sample size due to time limitations in EEG sessions. It is essential to understand the importance of the features in terms of physical electrode channels in order to design a high performance yet realistic BCI. The VSS produces a ranked list of original variables (electrode channels or features), according to their ability to discriminate between tasks. A linear discrimination analysis (LDA) classifier is applied to the selected variable subset, Evaluation of the VSS method using synthetic datasets selected more than 83% of relevant variables. Classification of imagery tasks using real BCI datasets resulted in less than 16% classification error.
| Original language | English (US) |
|---|---|
| Title of host publication | BIOSIGNALS 2009 - Proceedings of the 2nd International Conference on Bio-Inspired Systems and Signal Processing |
| Pages | 35-40 |
| Number of pages | 6 |
| State | Published - Jul 21 2009 |
| Event | 2nd International Conference on Bio-Inspired Systems and Signal Processing, BIOSIGNALS 2009 - Porto, Portugal Duration: Jan 14 2009 → Jan 17 2009 |
Publication series
| Name | BIOSIGNALS 2009 - Proceedings of the 2nd International Conference on Bio-Inspired Systems and Signal Processing |
|---|
Other
| Other | 2nd International Conference on Bio-Inspired Systems and Signal Processing, BIOSIGNALS 2009 |
|---|---|
| Country | Portugal |
| City | Porto |
| Period | 1/14/09 → 1/17/09 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Signal Processing
- Biomedical Engineering
- Control and Systems Engineering
Cite this
}
Variable subset selection for brain-computer interface : PCA-based dimensionality reduction and feature selection. / Dias, N. S.; Kamrunnahar, M.; Mendes, P. M.; Schiff, S. J.; Correia, J. H.
BIOSIGNALS 2009 - Proceedings of the 2nd International Conference on Bio-Inspired Systems and Signal Processing. 2009. p. 35-40 (BIOSIGNALS 2009 - Proceedings of the 2nd International Conference on Bio-Inspired Systems and Signal Processing).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Variable subset selection for brain-computer interface
T2 - PCA-based dimensionality reduction and feature selection
AU - Dias, N. S.
AU - Kamrunnahar, M.
AU - Mendes, P. M.
AU - Schiff, S. J.
AU - Correia, J. H.
PY - 2009/7/21
Y1 - 2009/7/21
N2 - A new formulation of principal component analysis (PCA) that considers group structure in the data is proposed as a Variable Subset Selection (VSS) method. Optimization of electrode channels is a key problem in brain-computer interfaces (BCI). BCI experiments generate large feature spaces compared to the sample size due to time limitations in EEG sessions. It is essential to understand the importance of the features in terms of physical electrode channels in order to design a high performance yet realistic BCI. The VSS produces a ranked list of original variables (electrode channels or features), according to their ability to discriminate between tasks. A linear discrimination analysis (LDA) classifier is applied to the selected variable subset, Evaluation of the VSS method using synthetic datasets selected more than 83% of relevant variables. Classification of imagery tasks using real BCI datasets resulted in less than 16% classification error.
AB - A new formulation of principal component analysis (PCA) that considers group structure in the data is proposed as a Variable Subset Selection (VSS) method. Optimization of electrode channels is a key problem in brain-computer interfaces (BCI). BCI experiments generate large feature spaces compared to the sample size due to time limitations in EEG sessions. It is essential to understand the importance of the features in terms of physical electrode channels in order to design a high performance yet realistic BCI. The VSS produces a ranked list of original variables (electrode channels or features), according to their ability to discriminate between tasks. A linear discrimination analysis (LDA) classifier is applied to the selected variable subset, Evaluation of the VSS method using synthetic datasets selected more than 83% of relevant variables. Classification of imagery tasks using real BCI datasets resulted in less than 16% classification error.
UR - http://www.scopus.com/inward/record.url?scp=67650544135&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=67650544135&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:67650544135
SN - 9789898111654
T3 - BIOSIGNALS 2009 - Proceedings of the 2nd International Conference on Bio-Inspired Systems and Signal Processing
SP - 35
EP - 40
BT - BIOSIGNALS 2009 - Proceedings of the 2nd International Conference on Bio-Inspired Systems and Signal Processing
ER -