ABSTRACT
This paper was aimed to investigate the changes of pulses under different emotional states.Based on the emotional pulse theory of traditional Chinese medicine (TCM),with college students as the research object,through films to stimulate different emotional expressions (i.e.,calm,sad and happy) of college students,the pulse recordings of college students were acquired in this study.The time domain parameters of pulses of college students under three types of emotional states were extracted.The difference among three groups of time-domain parameters was calculated by nonparametric test.Based on the time-domain parameters,the Support Vector Machine (SVM) classifier was used to classify three types of emotional states.The results showed that there were significant differences in three groups of time domain parameters (h3/h1,t and As/(As+Ad)).And the average recognition rate of three types of emotion states was 74.25％ based on the time-domain parameters of the pulse recordings.It was concluded that the pulse parameters can provide objective reference for emotional changes.It can be used to identify emotional states.