群体研讨支持系统中研讨主题的自动可视化聚类研究

群体研讨支持系统(Group Argument Support Systems, GASS)的匿名、并行输入及自动化记录群体发言的特征,在辅助群体产生大量有价值观点的同时,也常常导致"信息过载"和"知识断层".介绍了一个自动化聚类工具来增强群体的认知能力并提高电子会议的效率.首先识别了GASS环境下自动化主题聚类的一些挑战并回顾了相关研究,结合GASS的研讨模式、研讨文本特征及中文文本分析的要求,给出了中文分词、停词表处理以及有效词语识别的文本分析技术.提出基于主题分析的特征向量选择方法,并基于自组织映射的神经网络思想,用Java语言设计并开发了一个自动聚类工具.实验表明,该工具可以达到0.28的聚类准确率,0.35的聚类全面率,产生0.83的聚类错误率.

Research on Automatic Topic Visual Clustering in the Group Argument Support Systems

In Group Argument Support Systems (GASS), the characteristics of parallel and anonymous data entry often result in Information overload and knowledge breakout while improving group performance. This research presents an automatic topic clustering tool to enhance participants' cognitive ability and increase the productivity of electronic meetings. Firstly, we identify some challenges to automatic textual clustering and review the related research. Secondly, we describe a textual analysis method involving Chinese segment, stop list and effective terms identification. Thirdly, we propose a topic analysis method to obtain eigenvector. Fourthly, we design and implement an automatic clustering tool based on the self-Organization Map by Java. Experiment indicated that the tool can achieve clustering recall of 0.35, clustering precision of 0.28 and clustering error of 0.83. Finally, we give some discussions and future research directions.