课堂管理是衡量教师能力与课堂质量的重要指标之一，但现有的课堂管理研究缺乏有效的自动处理工具，难以面向庞大的教育体系进行大规模的研究分析。为了解决这一问题，本文运用基于规则的文本分类技术，提出并设计了一个基于师生对话文本的课堂管理行为自动分类器。 首先，本文研究并确定了课堂管理行为的分类标准；然后构建师生对话文本语料库，验证基于对话文本进行课堂管理行为研究的有效性；接着基于特征词和词性模板构造出了分类规则模板；最后，本文实现了一个课堂管理行为的自动分类器，可基于规则对对话文本进行自动统计与分类。实验结果表明：与传统的人工分类相比，该分类工具不仅能做到全面、准确，还能大幅减少研究人员的工作负担，提高研究效率，为相关研究提供技术支持。

Classroom management is one of the vital indicators to evaluate teaching quality. However, it’s hard to carry out large-scale analysis of it so far, for lack of automatic processing tool. Therefore, this paper designs an automatic classifier of classroom management behaviors based on the Teacher-Students dialogue texts, by use of rule based classification technique. First of all, this paper determines the classification criteria of classroom management behavior, then constructs the Teacher-Students dialogue text corpus and verifies the validity of using it to research. After that, the paper establishes classification rule templates, and finally implements an automatic classifier. Experiments show that: compared with the traditional artificial classification, this classification tool can not only keep comprehensive and accurate, but also can reduce the workload of the researchers heavily, improve the research efficiency and provide technical support for the related research.