Hidden Markov Models to Capture Sequential Patterns of Valence-Arousal in High- and Low-Performing Collaborative Problem-Solving Groups

Yaping Xu, Honghui Li, Weitong Guo, Tian Feng, Xiaonan Yin, Sen Bao, Lu Chen

2025

Abstract

Emotion is an important factor affecting students' cognitive processing and learning outcomes. Accurately detecting the group members’ emotions in collaborative problem-solving environments is an important basis for judging their learning status and providing personalized support. However, current research mainly focuses on discrete emotions and lacks the identification and analysis of learning emotions from the perspective of dimensional emotions, which may lead to an oversimplified representation of students' emotions. Therefore, based on the circumplex model of affect, this study used multiple machine learning methods to predict students' affective valence and arousal from facial behavioural clues when they participated in online collaborative problem-solving activities. The results indicated that the random forest model performed best. In order to enhance the understanding of the temporal nature of group emotions and their relationship with CPS outcomes, we also applied hidden Markov models (HMMs) to reveal the differences in sequential patterns between high- and low-performing groups. It was found that the sequential patterns of affective valence-arousal in the two groups of students were quite different, and students in the high-performing groups were more likely to regulate their emotions and transition to appropriate states (such as states with positive valence or high arousal) to successfully solve problems. This study has important methodological significance for the automatic measurement and analysis of dimensional emotions.

Paper Citation

in Harvard Style

Xu Y., Li H., Guo W., Feng T., Yin X., Bao S. and Chen L. (2025). Hidden Markov Models to Capture Sequential Patterns of Valence-Arousal in High- and Low-Performing Collaborative Problem-Solving Groups. In Proceedings of the 17th International Conference on Computer Supported Education - Volume 1: CSEDU; ISBN 978-989-758-746-7, SciTePress, pages 173-179. DOI: 10.5220/0013038900003932

in Bibtex Style

@conference{csedu25,
author={Yaping Xu and Honghui Li and Weitong Guo and Tian Feng and Xiaonan Yin and Sen Bao and Lu Chen},
title={Hidden Markov Models to Capture Sequential Patterns of Valence-Arousal in High- and Low-Performing Collaborative Problem-Solving Groups},
booktitle={Proceedings of the 17th International Conference on Computer Supported Education - Volume 1: CSEDU},
year={2025},
pages={173-179},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013038900003932},
isbn={978-989-758-746-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Computer Supported Education - Volume 1: CSEDU
TI - Hidden Markov Models to Capture Sequential Patterns of Valence-Arousal in High- and Low-Performing Collaborative Problem-Solving Groups
SN - 978-989-758-746-7
AU - Xu Y.
AU - Li H.
AU - Guo W.
AU - Feng T.
AU - Yin X.
AU - Bao S.
AU - Chen L.
PY - 2025
SP - 173
EP - 179
DO - 10.5220/0013038900003932
PB - SciTePress