Improving Memory Accesses for Heterogeneous Parallel Multi-objective Feature Selection on EEG Classification.
Juan José EscobarJulio OrtegaJesús GonzálezMiguel DamasPublished in: Euro-Par Workshops (2016)
Keyphrases
- feature selection
- multi objective
- classification accuracy
- classification models
- feature set
- support vector machine
- support vector
- machine learning
- text classification
- model selection
- evolutionary algorithm
- text categorization
- feature space
- high dimensionality
- motor imagery
- feature ranking
- feature subset
- discriminative features
- irrelevant features
- pattern recognition
- method for feature selection
- feature extraction
- multi class
- multi objective optimization
- eeg signals
- feature selection algorithms
- selecting relevant features
- parallel processing
- svm classifier
- genetic algorithm
- training set
- particle swarm optimization
- naive bayes
- optimization algorithm
- support vector machine svm
- image classification
- random access
- class labels
- objective function
- eeg data
- feature reduction
- class separability
- training data
- classification performances
- memory requirements
- sleep stage
- neural network