中文信息处理句法层面的研究是前一阶段的重点,而语义层面的研究尚处于起步阶段,但它对计算机语言理解来说更为关键;就一个句子的语义来说,非核心框架元素涵盖事件发生的时间、地点、方式等信息,在信息检索、信息提取和问答系统等应用领域是解决问题的重要依据.本文即根据中文框架元素实现规律的分析,研究非核心框架元素标注规则的构建方法.以认知语义领域介词结构的框架元素标注为例,构建了212条非核心框架元素标注规则,并通过实验分析了规则的使用效果和存在的问题,为计算机自动语义分析提供有效的方法.%Syntactic level research of Chinese information processing is the focus of the previous stage and semanticlevel research is still in its infancy, although it is even more critical for computational language understanding. Concerning one sentence' s meaning, the non-core frame elements often carry such information as events' time, place and manner and so on, which is the key to resolve the problems in such application fields as information searching, information extraction and question answering system, etc.. This paper thus focuses on the non-core frame elements recognition, based on the analysis of Chinese frame elements realization patterns. It obtains 212 non-core frame elements recognition rules, instantiating of preposition phrases in the cognitive semantics domain. It also analyzes the effects and problems of the rules through a test, providing an effective method for automatic semantic analysis.
展开▼
