Adaptive semi-supervised classification to reduce intersession non-stationarity in multiclass motor imagery-based brain-computer interfaces.
Luis F. Nicolás-AlonsoRebeca CorralejoJavier Gomez-PilarDaniel ÁlvarezRoberto HorneroPublished in: Neurocomputing (2015)
Keyphrases
- multi class
- brain computer interface
- motor imagery
- semi supervised classification
- semi supervised learning
- semi supervised
- healthy subjects
- spinal cord injury
- pairwise
- eeg signals
- labeled data
- unlabeled data
- signal processing
- cost sensitive
- binary classification
- support vector machine
- eeg data
- multi task
- binary classifiers
- brain activity
- non stationary
- data mining
- knn
- classification accuracy
- image processing
- feature selection
- computer vision