Emotion classification of teaching resources in online teaching based on artificial intelligence

To comprehensively analyze online course review texts and uncover the topics and emotion attitudes of learners, this study aims to enhance the quality of online teaching. This study focuses on analyzing learners’ discussion data from online courses and introduces an emotion dictionary specifically tailored for the MOOC (Massive Open Online Course) domain. The dictionary is constructed using a linear fusion algorithm that integrates SO-PMI (Semi-Supervised Offset Pointwise Mutual Information) with Word2Vec techniques. Additionally, the study proposes an enhanced Latent Dirichlet Allocation (LDA) model for topic mining and utilizes Support Vector Machines (SVM) for sentiment classification. By analyzing the textual data generated by learners, the study aims to reveal their focal points and emotional tendencies in greater depth. The experimental results show that the W-LDA model outperforms traditional LDA models in terms of predictive accuracy for topic distribution. Specifically, the proportions of high, moderate, and low positive emotions were 15.26%, 21.98%, and 50.06%, respectively, which exceed those of negative and neutral emotions. Furthermore, the sentiment analysis method based on topic distribution effectively uncovers latent information within course review texts, providing valuable insights for improving the quality of online teaching.