Bagging and Boosting Approach for EEG Signals Classification in Brain- Computer Interface Systems

Ensemble learning algorithms such as AdaBoost and Bagging have been used as an active research area and it has been shown that the classification results are improved for several benchmarking data sets with mainly decision trees as their base classijiers. In this paper we apply these Meta learning techniques with classifiers such as decision trees and support vector machines for BCI applications. The data set is &om Graz University in BCI Competition 2003. The task is to classify EEG signals in order to translate a subject's intentions into a technical control signal to control the peripheral environment. We compare the individual classijiers with their ensemble counterparts and discuss the results.

Keywords:

Brain-Computer Interface , Multiple Classifier System , Bagging , Boosting , Adabooost , Classification