EEG emotion recognition based on LMD fuzzy entropy

In order to improve the accuracy and reliability of EEG emotion recognition, an algorithm based on local mean decomposition (LMD) fuzzy entropy is proposed, and support vector machine (SVM) classification is performed on this basis. The LMD decomposition of the rhythm signal of the emotional EEG signal is performed to filter out the useless components of the EEG signal, and the PF components selected by the local mean decomposition algorithm are used as the input vector of the fuzzy entropy algorithm to extract the EEG signal feature parameters. Maximum correlation minimum redundancy is used for feature selection. Finally, the feature matrix is fed into the SVM classifier for binary classification in terms of validity and arousal, respectively, to achieve EEG emotion recognition. The method achieves 91.87% classification accuracy on the DEAP dataset, which can extract EEG emotion features more effectively and provides a new method for the study of EEG emotion recognition.