AUTHOR=Dan Yufang , Tao Jianwen , Zhou Di 

TITLE=Multi-Model Adaptation Learning With Possibilistic Clustering Assumption for EEG-Based Emotion Recognition

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 16 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2022.855421

DOI=10.3389/fnins.2022.855421

ISSN=1662-453X

ABSTRACT=In the field of machine learning, graph-based semi-supervised learning (GSSL) has attracted more and more attention due to its intuitive and good learning performance for emotion recognition. However, one of the reasons affecting the performance of GSSL method is that the training data and test data need to be independently identically distribution (IID), each individual subject may present completely different encephalogram(EEG) patterns in the same scenario that result in the data will be non-IID. In addition, there has limited effort has been made on improving GSSL’s performance by reducing the influence of noise/outlier EEG-based patterns. To this end, we propose in this paper a novel clustering method based on structure risk minimization model, called a Multi-model adaptation method of possibilistic clustering assumption (MA-PCA) effectively minimize the influence from the noise/outlier samples based on different EEG-based data distribution in some Reproduced Kernel Hilbert Space. Its main ideas are as follows: (1) reducing the negative influence of noise/outlier patterns for the method through fuzzy entropy regularization; (2) considering the training data and test data at IID and non-IID by exploiting the proposed multi-mode adaptive learning, and then obtain a better performance; (3) the algorithm implementation and convergence theorem also are given. A large number of experiments and analysis deeply on multiple real datasets (i.e., DEAP, SEED and SEED-IV) show that the proposed method has superior or comparable robustness and generalization performance of the EEG-based emotion recognition.