Study on BiliBili Curriculum Content Optimization Based on NLP

Abstract

With the rapid development of online education platforms, BiliBili has become an important channel for business learners to acquire knowledge, but its course content optimization still faces challenges such as dynamic changes in user needs and insufficient feedback processing efficiency. Based on natural language processing (NLP) technology, this study proposes a data-driven course content optimization framework, which aims to deeply analyze 20,000 user comments on BiliBili business courses through sentiment analysis, topic modeling, and keyword priority calculation. The results show that users are generally satisfied with the platform, with positive comments accounting for 64.7%, negative feedback 20.5%, and the remaining 14.8% reflecting neutral sentiments, which focuses on core issues such as difficulty in obtaining course resources, insufficient content practicality, and low interaction efficiency. Topic modeling further reveals that negative emotions are associated with resource and practicality disputes, neutral emotions reflect functional participation behaviors, and positive emotions focus on course depth and instructor professionalism. Based on sentiment-topic correlation analysis, this study proposes a priority-oriented optimization strategy, including a dynamic resource update mechanism, intelligent question-answering system development, and multimodal data integration, such as user click paths and video viewing behaviors, to improve the platform service efficiency. The research innovation lies in building an interdisciplinary evaluation model, addressing the limitations of traditional subjective feedback, and revealing users' dual needs for knowledge density and emotional value. In the future, it can be expanded to multidisciplinary scenarios, combined with educational psychology to deepen semantic understanding, explore dynamic recommendation systems, and promote the intelligent upgrade of the online education ecosystem.