计算机工程与应用 ›› 2021, Vol. 57 ›› Issue (11): 168-172.DOI: 10.3778/j.issn.1002-8331.2003-0199

• 模式识别与人工智能 • 上一篇    下一篇

基于灰色关联法的生理信号与情绪关联度研究

车敏诗,聂春燕,杨承金,阮新磊,范如俊   

  1. 长春大学 电子信息工程学院,长春 130022
  • 出版日期:2021-06-01 发布日期:2021-05-31

Research on Correlation Between Physiological Signal and Emotion Based on Grey Relational Method

CHE Minshi, NIE Chunyan, YANG Chengjin, RUAN Xinlei, FAN Rujun   

  1. School of Electronics and Information Engineering, Changchun University, Changchun 130022, China
  • Online:2021-06-01 Published:2021-05-31

摘要:

明确生理信号与情绪的关联度对提高情绪识别正确率起重要作用,然而目前关于两者的关联度研究成果比较少。为研究生理信号与情绪的关联度,采用德国Augsburg大学生理信号数据库的数据,基于灰色关联法研究喜、怒、哀情绪与心电、呼吸、皮肤电导信号的关联度,在此基础上,根据关联度结果采用CHAID决策树和SVM分类法进行情绪识别与分析。研究结果表明:(1)喜、怒、哀3种情绪与呼吸信号关联度最高,与皮肤电导关联度次之,与心电关联度最低;(2)基于CHAID决策树和SVM对3种情绪下的3种生理信号进行情绪识别,验证了喜、怒、哀与心电、呼吸、皮肤电导信号的关联程度。

关键词: 情绪, 生理信号, 关联度, 灰色关联分析法

Abstract:

The associations between physiological signals and emotions significantly improve the accuracy of emotion recognition, however, there are few achievements on them. To explore them, the paper adopts the physiological signal database?of Augsburg University(Germany) to research the correlation between emotions(joy, ?anger, sadness) among ECG, Respiration(RSP) and Skin Conductance(SC) on the basis of grey relational method, and then recognizes and analyzes emotions by CHAID Decision Tree(CHAID-DT) and SVM according to the relevancy findings. The results show that(1)joy, anger and sadness have the closest relationship with RSP, with SC the second and ECG the least; (2)it verifies the correlations between joy, anger and sadness among ECG, RSP and SC via recognizing three emotions with their three kinds of physiological signals by CHAID-DT and SVM.

Key words: emotion, physiological signal, correlation degree, grey relational analysis