Ensemble Learning in BCI-SSVEP Systems for Short Window Lengths
ResumoDifferent approaches have been investigated to implement effective Brain-Computer Interfaces (BCI), translating brain activation patterns into commands to external devices. BCI exploring Steady-State Visually Evoked Potentials usually achieve relatively high accuracy, when considering 2-3 second sample windows, but the performance degrades for smaller windows. So, we investigate the use of an ensemble method, the Adaboost algorithm, combining two different structures, the Logistic Regressor and the Multilayer Perceptron, whose diversity shall bring relevant information for more accurate classification. Simulation results indicate that the proposed method can improve performance for smaller windows, being a promising alternative to reduce model variance.
da Silva Jr., J. I. (2017). Comparativo de desempenho de sistemas bci-ssvep off-line e em tempo de execução utilizando técnicas de estimação de espectro e análise de correlação canônica. Master’s thesis, UFABC.
Freund, Y., Schapire, R., and Abe, N. (1999). A short introduction to boosting. JournalJapanese Society For Articial Intelligence, 14(771-780):1612.
Hastie, T., Rosset, S., Zhu, J., and Zou, H. (2009). Multi-class adaboost. Statistics and its Interface, 2(3):349–360.
Nam, C. S., Nijholt, A., and Lotte, F. (2018). Brain–Computer Interfaces Handbook: Technological and Theoretical Advances. CRC Press.