Adaptive classification helps hybrid visual brain computer interface systems handle non‐stationary cortical signals

Abstract The classifier efficiency of the brain‐computer interface systems is significantly impacted by non‐stationarity electroencephalogram (EEG) signals. We propose an adaptive variant linear discriminant analysis (LDA) as a solution to this problem. This constantly adjusts its parameters account for most recent EEG data. In study, authors will update mean values well covariance matrix each class pair. Visually evoked cortical potential datasets are used check how proposed performs. prove that LDA performs much better than both static multiclass and PMean LDA.