Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer Learning
For solving the problem of the inevitable decline in the accuracy of cross-subject emotion recognition via Electroencephalograph (EEG) signal transfer learning due to the negative transfer of data in the source domain, this paper offers a new method to dynamically select the data suitable for transf...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10017294/ |
| _version_ | 1797805049480478720 |
|---|---|
| author | Yuliang Ma Weicheng Zhao Ming Meng Qizhong Zhang Qingshan She Jianhai Zhang |
| author_facet | Yuliang Ma Weicheng Zhao Ming Meng Qizhong Zhang Qingshan She Jianhai Zhang |
| author_sort | Yuliang Ma |
| collection | DOAJ |
| description | For solving the problem of the inevitable decline in the accuracy of cross-subject emotion recognition via Electroencephalograph (EEG) signal transfer learning due to the negative transfer of data in the source domain, this paper offers a new method to dynamically select the data suitable for transfer learning and eliminate the data that may lead to negative transfer. The method which is called cross-subject source domain selection (CSDS) consists of the next three parts. 1) First, a Frank-copula model is established according to Copula function theory to study the correlation between the source domain and the target domain, which is described by the Kendall correlation coefficient. 2) The calculation method for the Maximum Mean Discrepancy is improved to determine the distance between classes in a single source. After normalization, the Kendall correlation coefficient is superimposed, and the threshold is set to identify the source-domain data most suitable for transfer learning. 3) In the process of transfer learning, on the basis of Manifold Embedded Distribution Alignment, the Local Tangent Space Alignment method is used to provide a low-dimensional linear estimation of the local geometry of nonlinear manifolds, which maintains the local characteristics of the sample data after dimensionality reduction. Experimental results show that compared with the traditional methods, the CSDS increases the accuracy of emotion classification by approximately 2.8% and reduces the runtime by approximately 65%. |
| first_indexed | 2024-03-13T05:46:11Z |
| format | Article |
| id | doaj.art-d104f5438938402e8a70bbb4ad91dbf6 |
| institution | Directory Open Access Journal |
| issn | 1558-0210 |
| language | English |
| last_indexed | 2024-03-13T05:46:11Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Transactions on Neural Systems and Rehabilitation Engineering |
| spelling | doaj.art-d104f5438938402e8a70bbb4ad91dbf62023-06-13T20:10:33ZengIEEEIEEE Transactions on Neural Systems and Rehabilitation Engineering1558-02102023-01-013193694310.1109/TNSRE.2023.323668710017294Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer LearningYuliang Ma0https://orcid.org/0000-0003-1277-4663Weicheng Zhao1https://orcid.org/0000-0001-9072-9740Ming Meng2Qizhong Zhang3Qingshan She4https://orcid.org/0000-0001-5206-9833Jianhai Zhang5School of Automation, Hangzhou Dianzi University, Hangzhou, ChinaSchool of HDU-ITMO Joint Institute, Hangzhou Dianzi University, Hangzhou, ChinaSchool of Automation, Hangzhou Dianzi University, Hangzhou, ChinaSchool of Automation, Hangzhou Dianzi University, Hangzhou, ChinaSchool of Automation, Hangzhou Dianzi University, Hangzhou, ChinaCollege of Computer Science and Technology, Hangzhou Dianzi University, Hangzhou, ChinaFor solving the problem of the inevitable decline in the accuracy of cross-subject emotion recognition via Electroencephalograph (EEG) signal transfer learning due to the negative transfer of data in the source domain, this paper offers a new method to dynamically select the data suitable for transfer learning and eliminate the data that may lead to negative transfer. The method which is called cross-subject source domain selection (CSDS) consists of the next three parts. 1) First, a Frank-copula model is established according to Copula function theory to study the correlation between the source domain and the target domain, which is described by the Kendall correlation coefficient. 2) The calculation method for the Maximum Mean Discrepancy is improved to determine the distance between classes in a single source. After normalization, the Kendall correlation coefficient is superimposed, and the threshold is set to identify the source-domain data most suitable for transfer learning. 3) In the process of transfer learning, on the basis of Manifold Embedded Distribution Alignment, the Local Tangent Space Alignment method is used to provide a low-dimensional linear estimation of the local geometry of nonlinear manifolds, which maintains the local characteristics of the sample data after dimensionality reduction. Experimental results show that compared with the traditional methods, the CSDS increases the accuracy of emotion classification by approximately 2.8% and reduces the runtime by approximately 65%.https://ieeexplore.ieee.org/document/10017294/Copula functionelectroencephalograph (EEG)emotion recognitionlocal tangent space alignment (LTSA)transfer learning |
| spellingShingle | Yuliang Ma Weicheng Zhao Ming Meng Qizhong Zhang Qingshan She Jianhai Zhang Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer Learning IEEE Transactions on Neural Systems and Rehabilitation Engineering Copula function electroencephalograph (EEG) emotion recognition local tangent space alignment (LTSA) transfer learning |
| title | Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer Learning |
| title_full | Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer Learning |
| title_fullStr | Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer Learning |
| title_full_unstemmed | Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer Learning |
| title_short | Cross-Subject Emotion Recognition Based on Domain Similarity of EEG Signal Transfer Learning |
| title_sort | cross subject emotion recognition based on domain similarity of eeg signal transfer learning |
| topic | Copula function electroencephalograph (EEG) emotion recognition local tangent space alignment (LTSA) transfer learning |
| url | https://ieeexplore.ieee.org/document/10017294/ |
| work_keys_str_mv | AT yuliangma crosssubjectemotionrecognitionbasedondomainsimilarityofeegsignaltransferlearning AT weichengzhao crosssubjectemotionrecognitionbasedondomainsimilarityofeegsignaltransferlearning AT mingmeng crosssubjectemotionrecognitionbasedondomainsimilarityofeegsignaltransferlearning AT qizhongzhang crosssubjectemotionrecognitionbasedondomainsimilarityofeegsignaltransferlearning AT qingshanshe crosssubjectemotionrecognitionbasedondomainsimilarityofeegsignaltransferlearning AT jianhaizhang crosssubjectemotionrecognitionbasedondomainsimilarityofeegsignaltransferlearning |