基于潜在语义分析的文本连贯性分析  

LATENT SEMANTIC ANALYSIS BASED TEXT COHERENCE ANALYSIS

作　　者：汤世平[1] 樊孝忠[1] 朱建勇[1] 

机构地区：[1]北京理工大学计算机科学技术学院,北京100081

出　　处：《计算机应用与软件》2008年第2期95-96,共2页Computer Applications and Software

基　　金：博士点基金(20050007023)

摘　　要：文本连贯性分析是计算机辅助评估中的重要内容,是对文本的表达质量进行评估的基础。提出一种将潜在语义分析方法与基于有序聚类的层次分析方法相结合的文本层次结构分析方法,该方法保证了层次划分的有序性,可操作性强,不依赖于具体领域。基本思想是:对于输入文本,首先识别文本物理结构,然后将文本依据主题划分为若干层次,最终获得文本的逻辑结构。实验结果表明,该方法是有效的,其准确率达到74.96%。Text coherence analysis is the key to computer assisted assessment. A hierarchical analysis method for text structure analysis is proposed. The method combines latent semantic analysis and hierarchical analysis method based on orderly cluster. The main idea is shown as follows：Firstly, the physical structure of text is obtained from the input text;Secondly, text hierarchical analysis is automatically performed;Finally,as a result,the logical structure of text is gained. The approach enables the hierarchies to keep paragraphs in a natural order,and it can be applied to many fields. The experimental result shows that the proposed method is effective, and it has achieved precision of 74.96%.

关 键 词：向量空间模型 潜在语义分析 文本连贯性 计算机辅助评估 

分 类 号：TP311.12[自动化与计算机技术—计算机软件与理论] H315[自动化与计算机技术—计算机科学与技术]