Multi-Output Regression for Integrated Prediction of Valence and Arousal in EEG-Based Emotion Recognition
This repository contains the code for the multi-output regression model used to predict the correlated dimensions of valence and arousal in EEG-based emotion recognition, aiming to improve prediction accuracy and efficiency. This project is based on the torcheeg framework.
- Single-output Regression: Predicts valence and arousal independently.
- Multi-output Regression: Predicts valence and arousal simultaneously, considering their interdependencies.
- Multi-output Regression with Chain Structure: Predicts valence first and uses it to predict arousal, reflecting the psychological sequence of emotional assessment.
DEAP_SD.ipynb
: Subject-Dependent projectGAMEEMO_SI.ipynb
: Subject-Independent project
@inproceedings{choi2024multi,
title={Multi-Output Regression for Integrated Prediction of Valence and Arousal in EEG-Based Emotion Recognition},
author={Choi, HyoSeon and Woo, ChaeEun and Kong, JiYun and Kim, Byung Hyung},
booktitle={2024 12th International Winter Conference on Brain-Computer Interface (BCI)},
pages={1--6},
year={2024},
organization={IEEE}
}
This repository has a MIT license, as found in the LICENSE file.
For any questions or issues, please contact HyoSeon Choi at [email protected].