Study of emotion recognition under anxiety based on physiological signals by relief method / 中国医疗器械杂志

ABSTRACT

Anxiety is usually generated because of the threatened feeling. The data of electrocardio, respiration, blood volume pulse and skin conductance signals were collected. The arithmetic of Relief were used for the feature selection and combined with k-Nearest Neighbor (kNN) arithmetic and Support Vector Machine (SVM) arithmetic for classification. The results show that the combination of Relief-SVM is better than combination of Relief-kNN on the recognition of anxiety state. The emotion recognition based on multi-physiological signals is superior to that based on one single signal.