Generalized sparse discriminant analysis for event-related potential classification

Victoria Peterson, Hugo Leonardo Rufiner, Ruben D. Spies

Abstract: Highlights•GSDA is a penalized version of the sparse discriminant analysis (SDA).•GSDA incorporates discriminative information by using Kullback–Leibler divergence.•An automatic method for the selection of regularization parameter is implemented.•The method is applied in small training size ERP-based BCI problems.•GSDA shows better performances compared to state-of-the-art methods.