Semi-Supervised EEG Clustering With Multiple Constraints
Chenglong Dai, Jia Wu, Jessica J. M. Monaghan, Guanghui Li, Hao Peng, Stefanie I. Becker, David McAlpine
Abstract
Electroencephalogram (EEG)-based applications in Brain-Computer Interfaces (BCIs, or Human-Machine Interfaces, HMIs), diagnosis of neurological disease, rehabilitation, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">etc</i> , rely on supervised techniques such as EEG classification that requires given class labels or markers. Incomplete or incorrectly labeled or unlabeled EEG data are increasing with the ever-expanding amount of EEG data generated by such applications and the ambiguities these generate degrade the performance of supervised techniques. To address the challenging task of clustering EEG data with limited <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">priori</i> knowledge, we introduce a semi-supervised graph embedding EEG clustering approach termed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ConsEEGc</i> with multiple constraints, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i.e.</i> , label-transformed connectivity constraints that constrains the connection or disconnection among EEG data, compactness-and-scatter constraint that constrains the intra-cluster compactness and inter-cluster scatter of EEG clusters, and fairness constraint that constrains the fair ratio of elements between EEG clusters, to make best use of limited <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">priori</i> knowledge of EEG data and to achieve better EEG clustering results. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ConsEEGc</i> is conducted with an optimization objective function that integrates pseudo label learning, least-square error minimization and multiple constraints, and it can quickly converge to local optima. The experiments demonstrate that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ConsEEGc</i> can efficiently yield good clustering results on various types of real-world EEG datasets, compared to state-of-the-art standard unsupervised and semi-supervised EEG/time series clustering algorithms.