Topic Mining of Public Opinion on Chemistry Textbooks Based on BERTopic Modeling: A Case Study on Zhihu Platform

Abstract

In the context of China's basic education textbook reform, chemistry textbooks have drawn increasing attention from all sectors of society. This study aims to identify the predominant themes in public comments on chemistry textbooks. This study uses the Zhihu platform as a data source and applies BERTopic topic modeling to identify 27 topics, from which six themes are derived through hierarchical clustering dendrograms. The results indicate that:(1) there are four tiers of needs—knowledge comprehension, resource supplementation, textbook use, and examination assessment.,(2) Public discourse on chemistry textbooks is not static but responsive to external policy changes, media controversies, and curriculum reform. Together, these findings point to the need for the textbook evaluation paradigm to shift from "static satisfaction assessment" to "dynamic usage behavior analysis," and the development of chemistry textbooks requires evidence from empirical analysis.