Group Nonnegative Matrix Factorization for EEG Classification

Hyekyoung Lee, Seungjin Choi
Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics, PMLR 5:320-327, 2009.

Abstract

Given EEG data measured from several subjects under the same condition, our goal is to estimate common task-related bases in a linear model that capture intra-subject variations as well as inter-subject variations. Such bases capture the common phenomenon in a group data, which is known as group analysis. In this paper we present a method of nonnegative matrix factorization (NMF) that is well suited to analyze EEG data of multiple subjects. The method is referred to as group nonnegative matrix factorization (GNMF) where we seek task-related common bases reflecting both intra-subject and inter-subject variations, as well as bases involving individual characteristics. We compare GNMF with NMF and some modified NMFs, in a task of learning spectral features from EEG data. Experiments on BCI competition data indicate that GNMF improves the EEG classification performance. In addition, we also show that GNMF is useful in a task of subject-to-subject transfer where the prediction for an unseen subject is performed based on a linear model learned from different subjects in the same group.

Cite this Paper

BibTeX
@InProceedings{pmlr-v5-lee09a, title = {Group Nonnegative Matrix Factorization for EEG Classification}, author = {Lee, Hyekyoung and Choi, Seungjin}, booktitle = {Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics}, pages = {320--327}, year = {2009}, editor = {van Dyk, David and Welling, Max}, volume = {5}, series = {Proceedings of Machine Learning Research}, address = {Hilton Clearwater Beach Resort, Clearwater Beach, Florida USA}, month = {16--18 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v5/lee09a/lee09a.pdf}, url = {https://proceedings.mlr.press/v5/lee09a.html}, abstract = {Given EEG data measured from several subjects under the same condition, our goal is to estimate common task-related bases in a linear model that capture intra-subject variations as well as inter-subject variations. Such bases capture the common phenomenon in a group data, which is known as group analysis. In this paper we present a method of nonnegative matrix factorization (NMF) that is well suited to analyze EEG data of multiple subjects. The method is referred to as group nonnegative matrix factorization (GNMF) where we seek task-related common bases reflecting both intra-subject and inter-subject variations, as well as bases involving individual characteristics. We compare GNMF with NMF and some modified NMFs, in a task of learning spectral features from EEG data. Experiments on BCI competition data indicate that GNMF improves the EEG classification performance. In addition, we also show that GNMF is useful in a task of subject-to-subject transfer where the prediction for an unseen subject is performed based on a linear model learned from different subjects in the same group.} }
Endnote
%0 Conference Paper %T Group Nonnegative Matrix Factorization for EEG Classification %A Hyekyoung Lee %A Seungjin Choi %B Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2009 %E David van Dyk %E Max Welling %F pmlr-v5-lee09a %I PMLR %P 320--327 %U https://proceedings.mlr.press/v5/lee09a.html %V 5 %X Given EEG data measured from several subjects under the same condition, our goal is to estimate common task-related bases in a linear model that capture intra-subject variations as well as inter-subject variations. Such bases capture the common phenomenon in a group data, which is known as group analysis. In this paper we present a method of nonnegative matrix factorization (NMF) that is well suited to analyze EEG data of multiple subjects. The method is referred to as group nonnegative matrix factorization (GNMF) where we seek task-related common bases reflecting both intra-subject and inter-subject variations, as well as bases involving individual characteristics. We compare GNMF with NMF and some modified NMFs, in a task of learning spectral features from EEG data. Experiments on BCI competition data indicate that GNMF improves the EEG classification performance. In addition, we also show that GNMF is useful in a task of subject-to-subject transfer where the prediction for an unseen subject is performed based on a linear model learned from different subjects in the same group.
RIS
TY - CPAPER TI - Group Nonnegative Matrix Factorization for EEG Classification AU - Hyekyoung Lee AU - Seungjin Choi BT - Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics DA - 2009/04/15 ED - David van Dyk ED - Max Welling ID - pmlr-v5-lee09a PB - PMLR DP - Proceedings of Machine Learning Research VL - 5 SP - 320 EP - 327 L1 - http://proceedings.mlr.press/v5/lee09a/lee09a.pdf UR - https://proceedings.mlr.press/v5/lee09a.html AB - Given EEG data measured from several subjects under the same condition, our goal is to estimate common task-related bases in a linear model that capture intra-subject variations as well as inter-subject variations. Such bases capture the common phenomenon in a group data, which is known as group analysis. In this paper we present a method of nonnegative matrix factorization (NMF) that is well suited to analyze EEG data of multiple subjects. The method is referred to as group nonnegative matrix factorization (GNMF) where we seek task-related common bases reflecting both intra-subject and inter-subject variations, as well as bases involving individual characteristics. We compare GNMF with NMF and some modified NMFs, in a task of learning spectral features from EEG data. Experiments on BCI competition data indicate that GNMF improves the EEG classification performance. In addition, we also show that GNMF is useful in a task of subject-to-subject transfer where the prediction for an unseen subject is performed based on a linear model learned from different subjects in the same group. ER -
APA
Lee, H. & Choi, S.. (2009). Group Nonnegative Matrix Factorization for EEG Classification. Proceedings of the Twelfth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 5:320-327 Available from https://proceedings.mlr.press/v5/lee09a.html.

Related Material