So sánh một số hàm phi tuyến trong thuật toán phân tích phần tử song song thích nghi cho ten-xơ bậc 3 và áp dụng xử lý tín hiệu điện não đồ không đầy đủ

Abstract: In this paper, we compared several nonlinear functions used in adaptive canonical polyadic (CP) decomposition of third-order tensors with one dimension growing with time. We also applied these adaptive CP decomposition algorithms to analysis of incomplete electroencephalogram (EEG) data under third-order tensor representation. The numerical results showed comparable performance of the algorithms when different nonlinear functions were used. The study also showed the applicability of adaptive canonical polyadic decomposition in tensor completion for EEG data.