Subclass discriminant analysis using dynamic cluster formation for EEG-based brain-computer interface

An important issue in pattern recognition is to extract features that have more discriminant power. In this context, one of the most effective approaches for optimal feature extraction is subclass discriminant analysis. However, in this approach, the major problem is to determine the optimal number of subclasses in each class. This paper presents a novel approach for efficient feature extraction via subclass discriminant analysis using dynamic cluster formation (DSDA). The main component of this paper is to design a method to automatically discover the optimal set of subclasses in each class. The effectiveness of the proposed method is evaluated by using the classification of EEG signals. The tasks to be discriminated are the imaginative hand movement and the resting state. The results demonstrate that the proposed method performed well in several experiments on different subjects and can improve the classification accuracy in the BCI systems.