A Kernel Partial Least Square Based Feature Selection Method
- Submitting institution
-
The University of Essex
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 1261
- Type
- D - Journal article
- DOI
-
10.1016/j.patcog.2018.05.012
- Title of journal
- Pattern Recognition
- Article number
- -
- First page
- 91
- Volume
- 83
- Issue
- -
- ISSN
- 0031-3203
- Open access status
- Compliant
- Month of publication
- May
- Year of publication
- 2018
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
2
- Research group(s)
-
B - Brain Computer Interfaces and Neural Engineering (BCI-NE)
- Citation count
- 9
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper proposes for the first-time a new relevance score function based on kernel partial least square regression for feature selection from high-dimensional data. It is significant because using the new relevance score in the state-of-the-art feature subset selection algorithm mRMR can achieve classification accuracy comparable to existing methods with significantly fewer features being selected, rendering real-time applications viable. This approach has been integrated in the online brain-computer-interfaces at Essex and featured in a recent review-article on BCI advances. ANOVA statistical tests based on extensive experiments (4 classifiers/7 datasets) using several classifiers and a number of datasets supported the findings.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -