Cluster impurity and forward-backward error maximization-based active learning for EEG signals classification.
Huijuan YangCuntai GuanKai Keng AngYaozhang PanHaihong ZhangPublished in: ICASSP (2012)
Keyphrases
- eeg signals
- active learning
- forward backward
- motor imagery
- signal processing
- brain computer interface
- eeg data
- text classification
- pattern recognition
- kernel density estimators
- fold cross validation
- brain activity
- pattern classification
- healthy subjects
- error rate
- image classification
- supervised learning
- classification accuracy
- support vector machine
- feature vectors
- selective sampling
- brain signals
- decision trees
- computer vision
- linear discriminant classifier
- electrical activity
- event related potentials
- machine learning
- extracted features
- cost sensitive
- decision rules
- hidden markov models
- training set
- feature space
- support vector
- feature extraction
- image segmentation
- clustering algorithm
- feature selection
- learning algorithm