语言模型在话题跟踪中的应用

语言模型技术作为信息检索领域的一个新的建模技术,已逐渐成为当代语言信息处理的主流技术之一。将该技术应用于话题跟踪研究中,对语言模型理论进行了介绍,详细描述了如何基于语言模型实现话题跟踪,构建了2个话题跟踪系统,分别利用向量空间模型和语言模型进行建模,并对它们的性能进行了比较。实验结果表明,语言模型比向量空间模型更适合于话题跟踪任务。

Application of Language Model Theory in Topic Tracking

Language model technology,as a new modeling technology in information retrieval field,has gradually become one of the essential technologies in current linguistic information processing.In the paper,this technology is used in topic tracking research.First,we introduce basics of language model theory,then,we describe how to apply language model theory to topic tracking.Finally,we set up two tracking systems,one based on language model theory and the other based on traditional VSM,to make performance comparison.Experimental results show that the proposed approach can achieve better performance than VSM based approach.