基于优化变分模态分解的脑电情绪识别

EMOTION RECOGNITION OF EEG BASED ON OPTIMAL VARIATIONAL MODE DECOMPOSITION

为提高脑电情绪识别的准确性与可靠性,提出一种基于优化变分模态分解(VMD)的脑电情绪识别方法。对情绪脑电的节律信号VMD分解,引入磷虾群优化算法(KH)搜索VMD的最优分解层数和惩罚因子;从分解后的固有模态分量(IMFs)中提取平均能量、功率谱密度作为特征;利用XGBoost算法进行分类。实验结果表明,与EMD、EEMD等特征提取方法相比,该方法在DEAP数据集上达到了91.02%的分类准确率,可以更有效地提取脑电情感特征,为脑电情绪识别的研究提供了新方法。

In order to improve the accuracy and reliability of EEG emotion recognition,a recognition method of EGG emotion based on optimal variational mode decomposition(VMD)is proposed.The rhythm signal of emotional EEG was decomposed by VMD,and the krill swarm optimization(KH)was introduced to search the optimal decomposition layer number and punishment factor of VMD.The average energy and power spectral density were extracted from the decomposed intrinsic modal component(IMFs)as features.The XGBoost algorithm was used for classification.The experimental results show that compared with EMD and EEMD,the classification accuracy of this method in DEAP dataset reaches 91.02%,which can more effectively extract EEG emotional features,and provide a new method for the study of EEG emotion recognition.