【作 者】：
杨晓哲， 王晴晴， 蒋佳龙
【关 键 词 】：
师生对话； 机器学习； 自动分类； 人工智能
【栏 目】：
课程与教学
【中文摘要】：
已有大量研究关注课堂中的师生对话，并对其进行了不同类型的编码与分析，但目前的编码仍受到专业人员的水平与时间限制。为提高课堂对话编码速度，实现自动化的课堂对话分类与即时反馈，采用人工智能技术，利用神经网络分析模型对课堂中的提问、回答、反馈进行自动编码。研究发现：该模型实现了基于语义的课堂对话质量评估，能够在短时间内实现课堂师生对话中IRE模型的分水平评估，即对提问层次水平、回答层次水平和反馈层次水平进行评估与分类；构建了全面、快速、准确的课堂对话评估方式，成为进一步理解课堂中学习发生过程的关键环节，为大规模课堂智能分析奠定重要基础。
【英文摘要】：
A large number of studies have focused on teacher-student dialogues in the classroom and different types of coding and analysis have been conducted for them, but the current coding is still limited by the level and time of professionals. In order to improve the coding speed of classroom dialogues and realize automatic classroom dialogue classification and instant feedback, a neural network analysis model is adopted to automatically encode questions, answers and feedback in class through artificial intelligence technology. The results show that the model realizes semantic-based quality assessment of classroom dialogues, and can realize the sub-level assessment of IRE model in classroom teacher-student dialogues in a short period of time, i.e., to assess and classify the questioning level, the answering level and the feedback level. This study constructs a comprehensive, fast and accurate classroom dialogue assessment method, which becomes a key link to further understand the learning process in the classroom, and lays an important foundation for large-scale classroom intelligent analysis.