在群体支持系统(GroupSupportSystems,GSS)的环境下,群体能够在很短时间内产生大量研讨文本,远远超过了人们对信息处理的能力。因此,迫切需要一种能够自动分析和处理群体研讨文本的方法,言语行为分类就是这类方法中有可能实现并且具有应用价值的一个。在分析Zeno研讨模型的基础上,提出了适合群体研讨语料的言语行为分类体系。采用基于转换学习的办法,通过引入多阶段转换学习的概念,初步解决了群体研讨文本言语行为分类的问题,并且在议题类别和一些表达主张的类别(如支持和反对)上取得了较好的识别效果。研究群体研讨文本的言语行为分类对于拓展GSS,进而研究和开发自动主持人系统具有重要意义。同时,也为在中文环境下解决其他类型研讨(如网络聊天室、即时聊天工具等)文本的言语行为分类问题提供了参考依据。
With the help of Group Support systems, group members may generate large amount of discus- sion texts in a short while, much more than what a human being can process. It is necessary provide an ap- proach to automatically analyze and process group discussion texts. Based on the analysis of Zeno argumen- tation model, we define a set of Speech Acts that are suitable for group discussion. By introducing multi- phase transformation based learning, we obtained satisfactory classification result for some Speech Act cat- egory such as positive proposition and negative proposition. This research is significant for extending Group Support Systems and developing automated facilitator for Group Support Systems in the future. It is also an example for future research on Speech Act classification in the Chinese context.