Modeling Employee Emotional Intelligence Progression Using Self-Supervised Deep Learning and Longitudinal Organizational Behavior Datasets

Emotional Intelligence (EI) plays a vital role in workplace dynamics, influencing communication, leadership, and overall organizational performance. Traditional methods for assessing EI progression rely on static surveys or supervised learning models, which struggle to capture the temporal complexity and evolution of emotions over time. These approaches often require large labeled datasets and lack generalization in dynamic, real-world environments. This paper proposes a novel self-supervised framework called EASTBiGRU (Emotion-Aware Self-supervised Temporal-BiGRU), designed to model EI progression using longitudinal organizational behavior datasets. The objective is to accurately predict EI scores and trends without depending heavily on annotated data. The model combines Temporal Convolutional Networks (TCN) to capture local and global temporal patterns with Bidirectional Gated Recurrent Units (BiGRU) for understanding emotional dynamics across time. Self-supervised pretraining enhances robustness by learning from unlabeled behavioral data, while the final fully connected layers provide score predictions. The proposed method outperforms recent models with a MAE of 3.5, RMSE of 4.2, and an <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$F_{1}$</tex> Score of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 5. 3 \%}$</tex>, demonstrating superior accuracy and stability. Tools used include Python, TensorFlow, Scikit-learn, and visualization libraries such as Matplotlib and Seaborn. This hybrid architecture offers a scalable and accurate solution for modeling emotional intelligence development in professional environments.

Перевод пока недоступен