Towards Context-aware EEG-based Emotion Recognition Models: Personality and Emotional Intelligence as Context

Emotion recognition is a critical component of affective computing technologies, enabling machines to understand and respond to human emotions more effectively. While traditional models rely on physiological signals, the inclusion of contextual factors, such as personality traits (PT) and emotional intelligence (EI), enhances the precision of these systems. In this paper, we propose a context-aware emotion recognition model using EEG signals, where PT and EI are integrated as additional inputs. Features were extracted using autoencoders, and models were tested with 64, 128, 256, and 512 feature sizes. XGBoost classifiers were employed for classification, and experiments were conducted in two phases: baseline models (using only EEG as input), and context-aware models (incorporating personality and EI scores along with EEG as input). Results show that context-aware models significantly outperformed baseline models, with the highest accuracy of 88.24% and F1-score of 0.8319 achieved when both contexts were included. Statistical tests confirmed a significant improvement in model performance with context, validating our hypothesis that context enhances emotion recognition accuracy. These findings highlight the importance of context awareness in advancing emotion recognition models for more accurate and reliable affective computing systems, with potential applications in mental health monitoring, personalized learning, and human-computer interaction.