Discussion on the Effect of Classroom Concept Learning Based on BERT Text Classification

In the process of concept learning, students will gradually construct concepts and eventually form a profound and complete concept system. Analyzing what students discuss in class can help teachers effectively understand students' level of conceptual learning and contribute to the development of teaching evaluation level. In this paper, we analyze students' conceptual learning levels by introducing the BERT combination model in deep learning. The research steps mainly include the introduction and formulation of concept learning classification metrics, the collection and preprocessing of datasets, and the construction of combinatorial optimization based on BERT models.Finally, the BERT-RCNN model achieved the best results, with an precision of 83.33%, a recall of 83.34%, and an F1-score of 83.34%.