多标签情感计算中的TSK模糊系统与域适应方法研究

何欣润; 李毅轩; 傅中正; 伍冬睿; 黄剑

doi:10.16383/j.aas.c240599

多标签情感计算中的TSK模糊系统与域适应方法研究

doi: 10.16383/j.aas.c240599 cstr: 32138.14.j.aas.c240599

何欣润^1,,
李毅轩^1,,
傅中正^1,,
伍冬睿^{1, 2,},
黄剑^1,

1.
华中科技大学人工智能与自动化学院类脑智能系统湖北省重点实验室武汉 430074
2.
深圳华中科技大学研究中心深圳 518000

基金项目: 国家重点研发计划(2022YFB4700200), 国家自然科学基金(62333007), 深圳市科技计划(JCYJ20220818103602004)资助

详细信息

作者简介:
何欣润：华中科技大学人工智能与自动化学院博士研究生. 2020年获得华中科技大学人工智能与自动化学院学士学位. 主要研究方向为机器学习, 情感计算, 多模态融合. E-mail: hexr@hust.edu.cn

李毅轩：华中科技大学人工智能与自动化学院硕士研究生. 2024年获得华中科技大学人工智能与自动化学院硕士学位. 主要研究方向为机器学习, 情感计算. E-mail: lyxuan@hust.edu.cn

傅中正：华中科技大学人工智能与自动化学院博士研究生. 主要研究方向为迁移学习, 情感计算, 手势识别. E-mail: fuzhongzheng@hust.edu.cn

伍冬睿：华中科技大学人工智能与自动化学院教授. 2003年获得中国科学技术大学自动控制学士学位, 2006年获得新加坡国立大学电气与计算机工程硕士学位, 2009年获得南加州大学洛杉矶分校电气工程博士学位. 主要研究方向为脑机接口, 机器学习, 计算智能, 情感计算. E-mail: drwu@hust.edu.cn

黄剑：华中科技大学人工智能与自动化学院教授. 1997年毕业于华中科技大学, 2000年获得华中科技大学工程硕士学位. 2005年获得华中科技大学博士学位. 2006年至2008年, 在日本名古屋大学微纳米系统工程系、机械信息与系统系担任博士后研究员. 主要研究方向为康复机器人, 机器人装配, 网络控制系统和生物信息学. 本文通信作者. E-mail: huang_jan@hust.edu.cn

计量
- 文章访问数: 34
- HTML全文浏览量: 24
- 被引次数: 0
出版历程
- 收稿日期: 2024-08-28
- 录用日期: 2025-03-02
- 网络出版日期: 2025-06-17

A study of TSK fuzzy system and domain adaptation method in multi-label affective computing

HE Xin-Run^1
,,
LI Yi-Xuan^1
,,
FU Zhong-Zheng^1
,,
WU Dong-Rui^{1, 2
,},
HUANG Jian^1
,

1.
School of Artificial Intelligence and Automation and Hubei Key Laboratory of Brain-inspired Intelligent Systems, Huazhong University of Science and Technology, Wuhan 430074
2.
Shenzhen Huazhong University of Science and Technology Research Institute, Shenzhen 518000

Funds: Supported by National Key R&D Program of China (2022YFB4700200), National Natural Science Foundation of China (62333007), Shenzhen Science and Technology Program (JCYJ20220818103602004)

More Information

Author Bio:
HE Xin-Run　Ph.D. candidate at the School of Artificial Intelligence and Automation, Huazhong University of Science and Technology. He received his bachelor degree from Huazhong University of Science and Technology in 2020. His research interest covers machine learning, affective computing, and multi-model fusion

LI Yi-Xuan　M. St. with the School of Artificial Intelligence and Automation, Huazhong University of Science and Technology. He received his master degree from Huazhong University of Science and Technology in 2024. His research interest covers machine learning and affective computing

FU Zhong-Zheng　Ph.D. candidate at the School of Artificial Intelligence and Automation, Huazhong University of Science and Technology. His research interest covers transfer learning, affective computing, and gesture recognition

WU Dong-Rui　Full professor with the School of Artificial Intelligence and Automation, Huazhong University of Science and Technology. He received a B.E in Automatic Control from the University of Science and Technology of China, Hefei, China, in 2003, an M.Eng in Electrical and Computer Engineering from the National University of Singapore in 2006, and a PhD in Electrical Engineering from the University of Southern California, Los Angeles, CA, in 2009. His research interest covers brain-computer interfaces, machine learning, computational intelligence, and affective computing

HUANG Jian　Full professor with the School of Artificial Intelligence and Automation, Huazhong University of Science and Technology (HUST). He graduated from HUST, China in 1997 and received the Master of Engineering degree from HUST in 2000. He received his Ph.D. from HUST in 2005. From 2006 to 2008, he was a postdoctoral researcher in the Department of Micro-Nano System Engineering and Department of Mechano-Informatics and Systems, Nagoya University, Japan. His main research interests include rehabilitation robot, robotic assembly, networked control systems and bioinformatics. Corresponding author of this paper

摘要

摘要: 情感计算作为人机交互领域的一个重要学科分支, 是实现和谐、自然的人机交互体验的关键保障. 如何利用便于获得的生理信号进行准确的情绪识别已成为了其中的热门话题. 广泛使用的情绪模型通常从愉悦维、唤醒维、支配维等多个维度描述情绪, 但现有情绪识别方法大多将不同维度分别考虑, 忽略了维度间的相关性关系, 并且在可解释性方面存在局限. 多标签TSK模糊系统虽然能够弥补以上不足, 但仍面临高维输入下模糊规则构建困难、训练效率低下的问题. 此外, 多模态生理信号具有较大的个体差异性, 严重影响了跨用户情绪识别的准确性. 本文首先提出了规则降维的多标签TSK模糊系统(RDR-MLTSK), 以优化模糊系统结构和训练效率; 进一步提出了多标签模糊域适应算法(MLFDA)实现多源域迁移学习, 提高了RDR-MLTSK的泛化性能. 在DEAP和DECAF两个公开数据集上的实验结果表明, 所提出的方法能有效提高情绪识别的准确率, 与经典和先进的方法相比具有更好的性能.
- 情感计算 /
- 多标签学习 /
- TSK模糊系统 /
- 领域自适应
Abstract: Affective computing, as an important branch of human-computer interaction (HCI), is a key guarantee for realizing a harmonious and natural HCI experience. How to utilize easily accessible physiological signals for accurate emotion recognition has become a hot topic. The widely used emotion model usually describes emotions from multiple dimensions, such as pleasure, arousal, and dominance, etc. However, most of the existing emotion recognition methods consider different dimensions separately, ignore the correlation relationship between dimensions, and have limitations in interpretability. Although the multi-label TSK fuzzy system can compensate for the above shortcomings, it still faces the problems of difficulty in constructing fuzzy rules and low training efficiency under high-dimensional input. In addition, multimodal physiological signals have large individual variability, which seriously affects the accuracy of cross-user emotion recognition. In this paper, we firstly propose a multi-label TSK fuzzy system with rule dimensionality reduction (RDR-MLTSK) to optimize the fuzzy system structure and the training efficiency; furthermore, we propose a multi-label fuzzy domain adaptation algorithm (MLFDA) to achieve multi-source domain migration learning, which improves the generalization performance of RDR-MLTSK. The experimental results on two publicly available datasets, DEAP and DECAF, show that the proposed methods can effectively improve the accuracy of emotion recognition and has better performance compared to classical and advanced methods.
- affective computing /
- multi-label learning /
- TSK fuzzy system /
- domain adaptation

HTML全文

图 1 SFCM模糊规则降维示意图

Fig. 1 SFCM fuzzy rule dimensionality reduction

下载: 全尺寸图片幻灯片

图 2 RDR-MLTSK模糊系统结构

Fig. 2 RDR-MLTSK fuzzy system structure

下载: 全尺寸图片幻灯片

图 3 MLFDA算法示意图

Fig. 3 MLFDA algorithm

下载: 全尺寸图片幻灯片

图 4 MLFDA算法流程框图

Fig. 4 MLFDA algorithm flow diagram

下载: 全尺寸图片幻灯片

图 5 多标签相关性分析结果

Fig. 5 Multi label correlation analysis results

下载: 全尺寸图片幻灯片

图 6 规则降维对准确率和训练时间的影响

Fig. 6 The impact of rule dimensionality reduction on accuracy and training time

下载: 全尺寸图片幻灯片

图 7 DEAP数据集各生理信号重要程度和参与程度

Fig. 7 The importance and involvement of various physiological signals in the DEAP dataset

下载: 全尺寸图片幻灯片

图 8 DECAF数据集各生理信号重要程度和参与程度

Fig. 8 The importance and involvement of various physiological signals in the DECAF dataset

下载: 全尺寸图片幻灯片

图 9 MLFDA算法准确率随$ K$变化曲线

Fig. 9 MLFDA algorithm accuracy versus K curve

下载: 全尺寸图片幻灯片

图 10 MLFDA算法准确率随训练迭代次数变化曲线

Fig. 10 The accuracy of MLFDA varies with training iterations

下载: 全尺寸图片幻灯片

图 11 K在不同取值时源域样本及聚类中心的t-SNE图

Fig. 11 t-SNE plots of source domain samples and cluster centers under different K values

下载: 全尺寸图片幻灯片

图 12 MLFDA算法准确率随$ m$变化曲线

Fig. 12 MLFDA algorithm accuracy versus $ m$ curve

下载: 全尺寸图片幻灯片

表 1 DEAP数据集特征提取

Table 1 Feature extraction from DEAP dataset

生理信号	提取的特征	特征维数
EEG	6个频带($ \delta $: 1 ~ 4 Hz, $ \theta $: 4 ~ 8 Hz, slow $ \alpha $: 8 ~ 10 Hz, $ \alpha $: 10 ~ 12 Hz, $ \beta $: 13 ~ 30 Hz, $ \gamma $:30-48Hz)的平均功率.	32通道×6维
EMG	均值, 最大值, 最小值, 标准差, 峰度, 偏度, 平均功率(20 ~ 64 Hz).	1通道×7维
GSR	均值, 标准差, 幅值, 峰值个数, 上升时间, 平均功率(0.4 ~ 1.0 Hz, 1.0 ~ 1.6 Hz, 1.6 ~ 2.2 Hz).	1通道×8维
BVP	心率的均值、最大值、最小值、标准差、均方根值, 心率变异性(R-R间期标准差).	1通道×6维

下载: 导出CSV

表 2 DECAF数据集特征提取

Table 2 Feature extraction from DECAF dataset

生理信号	提取的特征	特征维数
MEG	梯度计、磁强计筛选后通道的离散余弦变换的前8个系数.	9通道×8维×2
EMG	11个频带功率谱密度的6个统计特征(均值, 标准差, 峰度, 偏度, 标准分数大于1/小于−1的值所占比例).	1通道×66维
EOG	11个频带功率谱密度的6个统计特征(均值, 标准差, 峰度, 偏度, 标准分数大于1/小于−1的值所占比例).	1通道×66维
ECG	心率、心率变异性、心跳间隔、10个频带功率谱密度的统计特征等.	1通道×92维

下载: 导出CSV

表 3 DEAP数据集实验对比

Table 3 Experimental comparison on DEAP dataset

方法	愉悦维		唤醒维		支配维
方法	ACC	F1	ACC	F1	ACC	F1
LDA	0.620	0.532	0.604	0.598	0.584	0.571
SVM	0.632	0.596	0.612	0.603	0.575	0.564
LSTM	0.650	0.639	0.648	0.657	0.612	0.600
Koelstra等^[10]	0.630	0.610	0.620	0.580	0.602	0.589
Lian等^[62]	0.630	0.640	0.670	0.660	0.560	0.550
Kwade等^[63]	0.580	0.600	0.640	0.630	0.660	0.640
MLTSK	0.605	0.585	0.632	0.622	0.610	0.590
RDR-MLTSK	0.656	0.660	0.674	0.691	0.626	0.602

下载: 导出CSV

表 4 DECAF数据集实验对比

Table 4 Experimental comparison on DECAF dataset

方法	愉悦维		唤醒维		支配维
方法	ACC	F1	ACC	F1	ACC	F1
LDA	0.649	0.631	0.624	0.617	0.609	0.588
SVM	0.597	0.524	0.627	0.586	0.621	0.604
LSTM	0.673	0.679	0.688	0.667	0.652	0.640
Koelstra等^[10]	0.656	0.633	0.628	0.578	0.635	0.621
Lian等^[62]	0.686	0.664	0.696	0.681	0.634	0.605
Kwade等^[63]	0.645	0.620	0.684	0.673	0.658	0.647
MLTSK	0.644	0.658	0.643	0.629	0.631	0.609
RDR-MLTSK	0.690	0.706	0.680	0.689	0.671	0.659

下载: 导出CSV

表 5 直接迁移实验对比

Table 5 Comparison of direct transfer experiments

数据集	方法	愉悦维		唤醒维		支配维
数据集	方法	ACC	F1	ACC	F1	ACC	F1
DEAP	LSTM	0.650	0.639	0.648	0.657	0.612	0.600
	XGBoost	0.686	0.674	0.667	0.654	0.621	0.644
	He等^[67]	0.633	—	0.643	—	—	—
	Huang等^[68]	0.681	—	0.639	—	—	—
	Shen等^[69]	0.615	0.615	0.603	0.548	—	—
	MLTSK	0.609	0.635	0.618	0.631	0.594	0.626
	RDR-MLTSK	0.691	0.670	0.695	0.685	0.658	0.647
DECAF	LSTM	0.673	0.679	0.688	0.677	0.652	0.640
	XGBoost	0.707	0.684	0.687	0.666	0.651	0.654
	Zhang等^[70]	0.712	0.719	0.571	0.605	—	—
	MLTSK	0.629	0.611	0.634	0.627	0.619	0.585
	RDR-MLTSK	0.702	0.689	0.720	0.695	0.675	0.650
—: 文中未报告相应结果

下载: 导出CSV

表 6 不同相关性损失性能对比

Table 6 Performance comparison of different correlation loss

数据集	标签相关性	愉悦维		唤醒维		支配维
数据集	标签相关性	ACC	F1	ACC	F1	ACC	F1
DEAP	$ \alpha=0 $	0.659	0.623	0.676	0.643	0.616	0.614
	相关系数	0.691	0.670	0.695	0.685	0.658	0.647
	互信息	0.718	0.682	0.720	0.694	0.673	0.645
DECAF	$ \alpha=0 $	0.671	0.659	0.694	0.687	0.625	0.604
	相关系数	0.702	0.689	0.720	0.695	0.675	0.650
	互信息	0.693	0.682	0.711	0.684	0.681	0.664

下载: 导出CSV

表 7 DEAP数据集迁移学习实验对比

Table 7 Comparison of Transfer Learning Experiments on DEAP Dataset

方法	愉悦维		唤醒维		支配维
方法	ACC	F1	ACC	F1	ACC	F1
TCA	0.595	0.642	0.566	0.587	0.576	0.602
JDA	0.611	0.647	0.621	0.641	0.589	0.612
Fu等^[65]	0.636	0.653	0.644	0.679	0.611	0.623
Elalamy等^[66]	0.698	0.650	0.695	0.640	0.624	0.635
ML-SSJPDA(1NN)	0.636	0.653	0.644	0.675	0.590	0.621
ML-SSJPDA(SVM)	0.661	0.688	0.670	0.679	0.629	0.655
MLFDA	0.691	0.683	0.697	0.692	0.656	0.670

下载: 导出CSV

表 8 DECAF数据集迁移学习实验对比

Table 8 Comparison of Transfer Learning Experiments on DECAF Dataset

方法	愉悦维		唤醒维		支配维
方法	ACC	F1	ACC	F1	ACC	F1
TCA	0.622	0.603	0.623	0.597	0.606	0.614
JDA	0.651	0.626	0.642	0.606	0.629	0.642
Fu等^[65]	0.712	0.719	0.671	0.695	0.627	0.626
Elalamy等^[66]	0.610	0.610	0.600	0.600	0.621	0.611
ML-SSJPDA(1NN)	0.676	0.683	0.694	0.665	0.650	0.678
ML-SSJPDA(SVM)	0.711	0.688	0.670	0.679	0.629	0.655
MLFDA	0.736	0.698	0.742	0.724	0.698	0.696

下载: 导出CSV

表 9 多标签学习性能对比

Table 9 Comparison of multi-label learning performance

数据集	评价指标	CC	BRNN	MLKNN	MLFDA
DEAP	SA	0.41	0.43	0.37	0.45
DEAP	HL	0.31	0.29	0.36	0.26
DECAF	SA	0.52	0.52	0.44	0.54
DECAF	HL	0.19	0.22	0.28	0.14

下载: 导出CSV

参考文献(72)

[1]	Suhaimi N S, Mountstephens J, Teo J. EEG-based emotion recognition: a state-of-the-art review of current trends and opportunities. Computational Intelligence and Neuroscience, 2020, 2020(1): 8875426
[2]	Cambria E, Das D, Bandyopadhyay S, Feraco A. Affective computing and sentiment analysis. A Practical Guide to Sentiment Analysis, 20171−10
[3]	De Nadai S, D'Incà M, Parodi F, Benza M, Trotta A, Zero E, et al. Enhancing safety of transport by road by online monitoring of driver emotions. In: Proceedings of the 11th System of Systems Engineering Conference. IEEE, 2016. 1−4
[4]	Huang H Y, Xie Q Y, Pan J H, He Y B, Wen Z F, Yu R H, et al. An EEG-based brain computer interface for emotion recognition and its application in patients with disorder of consciousness. IEEE Transactions on Affective Computing, 2021, 12(4): 832−842 doi: 10.1109/TAFFC.2019.2901456
[5]	权学良, 曾志刚, 蒋建华, 张亚倩, 吕宝粮, 伍冬睿. 基于生理信号的情感计算研究综述. 自动化学报, 2021, 47(8): 1769−1784 Quan Xue-Liang, Zeng Zhi-Gang, Jiang Jian-Hua, Zhang Ya-Qian, Lv Bao-Liang, Wu Dong-Rui. Physiological signals based affective computing:a systematic review. Acta Automatica Sinica, 2021, 47(8): 1769−1784
[6]	Wu D R, Lu B L, Hu B, Zeng Z G. Affective Brain-Computer Interfaces (aBCIs): A Tutorial. Proceedings of the IEEE, 2023, 111(10): 1314−1332 doi: 10.1109/JPROC.2023.3277471
[7]	Zeng Z H, Pantic M, Roisman G I, Huang T S. A survey of affect recognition methods: audio, visual, and spontaneous expressions. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009, 31(1): 39−58 doi: 10.1109/TPAMI.2008.52
[8]	Arifin S, Cheung P Y K. Affective level video segmentation by utilizing the pleasure-arousal-dominance information. IEEE Transactions on Multimedia, 2008, 10(7): 1325−1341 doi: 10.1109/TMM.2008.2004911
[9]	Poria S, Cambria E, Bajpai R, Hussain A. A review of affective computing: From unimodal analysis to multimodal fusion. Information fusion, 2017, 37: 98−125 doi: 10.1016/j.inffus.2017.02.003
[10]	Koelstra S, Muhl C, Soleymani M, Lee J-S, Yazdani A, Ebrahimi T, et al. DEAP: A Database for Emotion Analysis Using Physiological Signals. IEEE Transactions on Affective Computing, 2012, 3(1): 18−31 doi: 10.1109/T-AFFC.2011.15
[11]	Abadi M K, Subramanian R, Kia S M, Avesani P, Patras I, Sebe N. DECAF: MEG-Based Multimodal Database for Decoding Affective Physiological Responses. IEEE Transactions on Affective Computing, 2015, 6(3): 209−222 doi: 10.1109/TAFFC.2015.2392932
[12]	Soleymani M, Lichtenauer J, Pun T, Pantic M. A multimodal database for affect recognition and implicit tagging. IEEE Transactions on Affective Computing, 2012, 3(1): 42−55 doi: 10.1109/T-AFFC.2011.25
[13]	Katsigiannis S, Ramzan N. DREAMER: A database for emotion recognition through EEG and ECG signals from wireless low-cost off-the-shelf devices. IEEE Journal of Biomedical and Health Informatics, 2018, 22(1): 98−107 doi: 10.1109/JBHI.2017.2688239
[14]	Alvarado N. Arousal and valence in the direct scaling of emotional response to film clips. Motivation and Emotion, 1997, 21: 323−348 doi: 10.1023/A:1024484306654
[15]	Lewis P A, Critchley H D, Rotshtein P, Dolan R J. Neural correlates of processing valence and arousal in affective words. Cerebral Cortex, 2007, 17(3): 742−748
[16]	Oliveira A M, Teixeira M P, Fonseca I B, Oliveira M. Joint Model-Parameter Validation of Self-Estimates of Valence and Arousal: Probing a Differential-Weighting Model of Affective Intensity. Proceedings of Fechner Day, 2006, 22(1): 245−250
[17]	Fourati R, Ammar B, Aouiti C, Sanchez-Medina J, Alimi A M. Optimized echo state network with intrinsic plasticity for EEGbased emotion recognition. In: Proceedings of the 24th International Conference on Neural Information Processing. Guangzhou, China: Springer, 2017. 718−727
[18]	Ren F J, Dong Y D, Wang W. Emotion recognition based on physiological signals using brain asymmetry index and echo state network. Neural Computing and Applications, 2019, 31(9): 4491−4501 doi: 10.1007/s00521-018-3664-1
[19]	Liu Y, Ding Y F, Li C, Cheng J, Song R C, Wan F, et al. Multi-channel EEG-based emotion recognition via a multi-level features guided capsule network. Computers in Biology and Medicine, 2020, 123: Article No. 103927 doi: 10.1016/j.compbiomed.2020.103927
[20]	Hou K, Zhang X W, Yang Y K, Zhao Q Q, Yuan W J, Zhou Z Y, et al. Emotion recognition from multimodal physiological signals via discriminative correlation fusion with a temporal alignment mechanism. IEEE Transactions on Cybernetics, 2024, 54(5): 3079−3092 doi: 10.1109/TCYB.2023.3320107
[21]	Zhang Z, Zhong S H, Liu Y. Beyond mimicking under-represented emotions: Deep data augmentation with emotional subspace constraints for EEG-based emotion recognition. Proceedings of the AAAI Conference on Artificial Intelligence, 2024, 38(9): 10252−10260 doi: 10.1609/aaai.v38i9.28891
[22]	Nicolaou M A, Gunes H, Pantic M. Continuous prediction of spontaneous affect from multiple cues and modalities in valence-arousal space. IEEE Transactions on Affective Computing, 2011, 2(2): 92−105 doi: 10.1109/T-AFFC.2011.9
[23]	Rivas J J, del Carmen Lara M, Castrejon L, Hernández-Franco J, Orihuela-Espina F, Palafox L, et al. Multi-Label and Multimodal Classifier for Affective States Recognition in Virtual Rehabilitation. IEEE Transactions on Affective Computing, 2022, 13(3): 1183−1194 doi: 10.1109/TAFFC.2021.3055790
[24]	Wu D R, Huang J. Affect Estimation in 3D Space Using Multi-Task Active Learning for Regression. IEEE Transactions on Affective Computing, 2022, 13(1): 16−27 doi: 10.1109/TAFFC.2019.2916040
[25]	Zheng W J, Yu J F, Xia R. A Unimodal Valence-Arousal Driven Contrastive Learning Framework for Multimodal Multi-Label Emotion Recognition. In: Proceedings of the ACM Multimedia. Melbourne, Australia: ACM, 2024. 1−10
[26]	Li X Y, Liu J, Xie Y R, Gong P Z, Zhang X L, He H H. MAGDRA: A Multi-modal Attention Graph Network with Dynamic Routing-By-Agreement for multi-label emotion recognition. Knowledge-Based Systems, 2024, 283: 111126 doi: 10.1016/j.knosys.2023.111126
[27]	Cui Y Q, Xu Y F, Wu D R. EEG-Based Driver Drowsiness Estimation Using Feature Weighted Episodic Training. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2019, 27(11): 2263−2273 doi: 10.1109/TNSRE.2019.2945794
[28]	Kar S, Das S, Ghosh P K. Applications of neuro fuzzy systems: A brief review and future outline. Applied Soft Computing, 2014, 15: 243−259 doi: 10.1016/j.asoc.2013.10.014
[29]	Lotte F, Bougrain L, Cichocki A, Clerc M, Congedo M, Rakotomamonjy A, et al. A review of classification algorithms for EEG-based brain-computer interfaces: A 10 year update. Journal of Neural Engineering, 2018, 15(3): 1005−1033
[30]	Lu W, Chen Y Q, Wang J D, Qin X. Cross-domain activity recognition via substructural optimal transport. Neurocomputing, 2021, 454(1): 65−75
[31]	Chai X, Wang Q S, Zhao Y P, Li Y Q, Liu D, Liu X, et al. A fast, efficient domain adaptation technique for cross-domain electroencephalography(EEG)-based emotion recognition. Sensors, 2017, 17(5): 1014−1034 doi: 10.3390/s17051014
[32]	Zhang W, Deng L F, Zhang L, Wu D R. A Survey on Negative Transfer. IEEE/CAA Journal of Automatica Sinica, 2023, 10(2): 305−329 doi: 10.1109/JAS.2022.106004
[33]	Li J P, Qiu S, Shen Y Y, Liu C L, He H G. Multisource transfer learning for cross-subject EEG emotion recognition. IEEE Transactions on Cybernetics, 2019, 50(7): 3281−3293
[34]	Lin Y P, Jung T P. Improving EEG-based emotion classification using conditional transfer learning. Frontiers in Human Neuroscience, 2017, 11(1): 334−344
[35]	Lou Q D, Deng Z H, Xiao Z Y, Choi K S, Wang S T. Multilabel Takagi-Sugeno-Kang Fuzzy System. IEEE Transactions on Fuzzy Systems, 2022, 30(9): 3410−3425 doi: 10.1109/TFUZZ.2021.3115967
[36]	Bian Z W, Chang Q, Wang J, Pedrycz W, Pal N R. Takagi-Sugeno-Kang Fuzzy Systems for High-Dimensional Multilabel Classification. IEEE Transactions on Fuzzy Systems, 2024, 32(6): 3790−3804 doi: 10.1109/TFUZZ.2024.3382981
[37]	Krinidis S, Chatzis V. A robust fuzzy local information C-means clustering algorithm. IEEE transactions on image processing, 2010, 19(5): 1328−1337 doi: 10.1109/TIP.2010.2040763
[38]	Cui Y Q, Wu D R, Xu Y F. Curse of Dimensionality for TSK Fuzzy Neural Networks: Explanation and Solutions. In: 2021 International Joint Conference on Neural Networks. Shenzhen, China: IEEE, 2021. 1−8
[39]	Houle M E, Kriegel H P, Kröger P, Schubert E, Zimek A. Can shared-neighbor distances defeat the curse of dimensionality?. In: Scientific and Statistical Database Management: 22nd International Conference. Heidelberg, Germany: Springer, 2010. 482−500
[40]	Winkler R, Klawonn F, Kruse R. Fuzzy c-means in high dimensional spaces. International Journal of Fuzzy System Applications, 2011, 1(1): 1−16 doi: 10.4018/ijfsa.2011010101
[41]	Urbanowicz R J, Meeker M, La Cava W, Olson R S, Moore J H. Relief-based feature selection: Introduction and review. Journal of Biomedical Informatics, 2018, 85: 189−203 doi: 10.1016/j.jbi.2018.07.014
[42]	Gu X M, Cheng X. Distilling a deep neural network into a TakagiSugeno-Kang fuzzy inference system. arxiv: 2010.04974, 2020
[43]	Du G L, Wang Z, Li C, Liu P X. A TSK-type convolutional recurrent fuzzy network for predicting driving fatigue. IEEE Transactions on Fuzzy Systems, 2020, 29(8): 2100−2111
[44]	Lin K P. A Novel Evolutionary Kernel Intuitionistic Fuzzy C-means Clustering Algorithm. IEEE Transactions on Fuzzy Systems, 2014, 22(5): 1074−1087 doi: 10.1109/TFUZZ.2013.2280141
[45]	Li Y X, Fu Z Z, He X R, Huang J. An optimized multi-label TSK fuzzy system for emotion recognition of multimodal physiological signals. In: 2022 IEEE International Conference on Cyborg and Bionic Systems. Wuhan, China: IEEE, 2023. 362−367
[46]	Jang J S R. ANFIS: adaptive-network-based fuzzy inference system. IEEE Transactions on Systems, Man, and Cybernetics, 1993, 23(3): 665−685 doi: 10.1109/21.256541
[47]	Wu D R, Yuan Y, Huang J, Tan Y H. Optimize TSK Fuzzy Systems for Regression Problems: Minibatch Gradient Descent With Regularization, DropRule, and AdaBound (MBGD-RDA). IEEE Transactions on Fuzzy Systems, 2019, 28(5): 1003−1015
[48]	汪雷, 黄剑, 段涛, 伍冬睿, 熊蔡华, 崔雨琦. 基于气压肌动图和改进神经模糊推理系统的手势识别研究. 自动化学报, 2022, 48(5): 1220−1233 Wang Lei, Huang Jian, Duan Tao, Wu Dong-Rui, Xiong Cai-Hua, Cui Yu-Qi. Research on Gesture Recognition Based on Pressure-based Mechanomyogram and Improved Neural Fuzzy Inference System. Acta Automatica Sinica, 2022, 48(5): 1220−1233
[49]	Wang J D, Chen Y Q, Feng W J, Yu H, Huang M Y, Yang Y. Transfer learning with dynamic distribution adaptation. ACM Transactions on intelligent Systems and Technology, 2020, 11(1): 1−25
[50]	Fu Z Z, Zhang B N, He X R, Li Y X, Wang H Y, Huang J. Emotion recognition based on multi-modal physiological signals and transfer learning. Frontiers in Neuroscience, 2022, 16(1): 716−730
[51]	Zhang W, Wu D R. Discriminative joint probability maximum mean discrepancy(DJP-MMD) for domain adaptation. In: 2020 International Joint Conference on Neural Networks. Glasgow, United Kindom: IEEE, 2020. 1−8
[52]	Pan S J, Tsang I W, Kwok J T, Yang Q. Domain adaptation via transfer component analysis. IEEE Transactions on Neural Networks, 2010, 22(2): 199−210
[53]	Ling M S, Wang J M, Ding G G, Sun J G, Yu P S. Transfer feature learning with joint distribution adaptation. In: Proceedings of the IEEE International Conference on Computer Vision. Sydney, Australia: IEEE, 2013. 2200–2207
[54]	Pan S J, Yang Q. A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering, 2009, 22(10): 1345−1359
[55]	Khan M A A H, Roy N, Misra A. Scaling human activity recognition via deep learning-based domain adaptation. In: 2018 IEEE International Conference on Pervasive Computing and Communications. Athens, Greece: IEEE, 2018. 1−9
[56]	Lu J, Zuo H, Zhang G Q. Fuzzy Multiple-source Transfer Learning. IEEE Transactions on Fuzzy Systems, 2020, 28(12): 3418−3431 doi: 10.1109/TFUZZ.2019.2952792
[57]	Li C H, Zhang R, Li J, Petre S. Bayesian Information Criterion for Signed Measurements with Application to Sinusoidal Signals. IEEE Signal Processing Letters, 2018, 25(8): 1251−1255 doi: 10.1109/LSP.2018.2847559
[58]	Read J, Pfahringer B, Holmes G, Frank E. Classifier chains for multi-label classification. Machine learning, 2011, 85(3): 333−359 doi: 10.1007/s10994-011-5256-5
[59]	Zhang M L, Zhou Z H. ML-KNN: A lazy learning approach to multi-label learning. Pattern Recognition, 2007, 40(7): 2038−2048 doi: 10.1016/j.patcog.2006.12.019
[60]	Spyromitros E, Tsoumakas G, Vlahavas I. An empirical study of lazy multilabel classification algorithms. In: Artificial Intelligence: Theories, Models and Applications: 5th Hellenic Conference on AI. Syros, Greece: Springer, 2008. 401−406
[61]	He H H, Xia R. Joint binary neural network for multi-label learning with applications to emotion classification. In: Natural Language Processing and Chinese Computing. Hohhot, China: Springer, 2018. 250−259
[62]	连卫芳, 晁浩, 刘永利. 基于SDAE和RELM的EEG情感识别方法. 计算机工程, 2021, 47(9): 75−83 Lian Wei-Fang, Chao Hao, Liu Yong-Li. EEG Emotion Recognition Method Based on SDAE and RELM. Computer Engineering, 2021, 47(9): 75−83
[63]	Kawde P, Verma G K. Deep belief network based affect recognition from physiological signals. In: 2017 4th IEEE Uttar Pradesh Section International Conference on Electrical, Computer and Electronics. Mathura, India: IEEE, 2017. 587−592
[64]	Pereira E T, Gomes H M, Veloso L R, Mota M R A. Empirical Evidence Relating EEG Signal Duration to Emotion Classification Performance. IEEE Transactions on Affective Computing, 2021, 12(1): 154−164 doi: 10.1109/TAFFC.2018.2854168
[65]	Fu Z Z, He X R, Wang E K, Huo J, Huang J, Wu D R. Personalized human activity recognition based on integrated wearable sensor and transfer learning. Sensors, 2021, 21(3): 885−908 doi: 10.3390/s21030885
[66]	Elalamy R, Fanourakis M, G. Chanel. Multi-modal emotion recognition using recurrence plots and transfer learning on physiological signals. In: 2021 9th International Conference on Affective Computing and Intelligent Interaction. Nara, Japan: IEEE, 2021. 1−7
[67]	He Z P, Zhong Y S, Pan J H. An adversarial discriminative temporal convolutional network for EEG-based cross-domain emotion recognition. Computers in Biology and Medicine, 2022, 141: 105048 doi: 10.1016/j.compbiomed.2021.105048
[68]	Huang D M, Chen S T, Liu C, Zheng L, Tian Z H, Jiang D Z. Differences first in asymmetric brain: A bi-hemisphere discrepancy convolutional neural network for EEG emotion recognition. Neurocomputing, 2021, 448: 140−151 doi: 10.1016/j.neucom.2021.03.105
[69]	Shen J, Zhu K X, Liu H K, Wu J W, Wang K, Dong Q X. Tensor Correlation Fusion for Multimodal Physiological Signal Emotion Recognition. IEEE Transactions on Computational Social Systems, 2024, early access
[70]	Zhang X W, Liu J Y, Shen J, Li S J, Hou K C, Hu B, et al. Emotion Recognition From Multimodal Physiological Signals Using a Regularized Deep Fusion of Kernel Machine. IEEE Transactions on Cybernetics, 2021, 51(9): 4386−4399 doi: 10.1109/TCYB.2020.2987575
[71]	Angelov P, Filev D P, Kasabov N. Evolving intelligent systems: methodology and applications. John Wiley & Sons, 2010
[72]	Zhou J, Chen L, Chen C P, Zhang Y, Li H X. Fuzzy clustering with the entropy of attribute weights. Neurocomputing, 2016, 198: 125−134 doi: 10.1016/j.neucom.2015.09.127