An Emotion Recognition with Online Learning Data Using Deep Learning Approaches

Emotion recognition helps teachers to obtain students' emotional states and their learning progress in their study. Many deep learning-based technologies, such as convolutional neural networks (CNN) and large language models (LLM) models are used for emotion recognition and prediction with learning data of students. An emotion recognition and prediction for online learning students using deep learning approaches is proposed, namely EPLLM. The method transforming the students' data, such as forum texts, questions, and answers, to a sequential form and extracts feature information from students' data using deep learning models with their behavioral changes in online learning. The student emotions are recognized and predicted based online learning data with combination of the deep learning models. To verify the performances of the model, many experiments are conducted using the public dataset. The results verify that, the model achieved better effectiveness in emotion prediction, compared with existing models. The values of F1 score of the proposed model is improved 0.051 compared that of the exist SGNN-LLM model.