项目名称: 汉语句法分析中的自动歧义识别和分类问题研究
项目编号: No.61300158
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 黄书剑
作者单位: 南京大学
项目金额: 23万元
中文摘要: 高效的信息处理应用需要有效的文本的自动分析和理解方法,句法分析是文本分析体系中的重要环节。句法分析效果在实际应用中仍然难以让人满意,这主要是由于对句法结构歧义的处理错误导致的。句法结构歧义是指存在相同或相似的句子片段对应多个不同的句法结构的情况。以往的研究大多关注于部分歧义实例的消解、某个具体的歧义问题或特定的语义资源的使用,缺乏通用的歧义处理手段。本项目拟研究在自动句法分析中的歧义识别方法以及面向消解过程的歧义分类方法。首先采用不确定性分析的技术对句法分析中影响分析效果的关键性歧义自动识别;然后结合语言学的歧义理论,根据消解这些歧义所需要的不同上下文和语义知识来源对歧义进行分类,为自动的歧义消解提供依据。
中文关键词: 句法分析;歧义识别;歧义消解;;
英文摘要: Efficient information processing applications often require effective methods to automatically analyze and understand the text. Syntactic analysis, or parsing, is an important step of the text analysis pipeline. However, the parsing result of real world applications is usually not good enough. An important factor of parsing errors is the structural ambiguity. Structural ambiguity refers to the presence of multiple syntactic structures for the same or similar sentence fragments. Most of the previous studies focus on the theoretical issues, the analysis of given ambiguous instances or the application of a certain linguistic resource. In this project we plan to study methods for automatically identifying and classifying structural ambiguities. We firstly identify structural ambiguities using uncertainty measures. Then we classify these ambiguities into different categories according to the context or linguistic resources requires to solve them. These identification and classification results may leads to specific resolution techniques for a certain ambiguity type and better resolution result.
英文关键词: Syntactic Parsing;Ambiguity Recognition;Ambiguity Resolution;;