机用义项库在词义标注中的作用及其完善

Modification and Improvement of Traditional Dictionary by WSB

计算机词义标注一般都沿用传统词典的义项,随着研究的深入,对传统词典义项的设立提出了新的要求。有学者提出了以句法功能为主要依据来辨析义项。我们的研究注重词义语义特征的提取,共建构了七个资源库,互相之间主要围绕着词义语义特征展开,义项库在计算机词义标注中居于关键位置。机用义项库与传统词典义项相比,在词义标注目的、对语料的覆盖力、义项间语义关系、义项的可识别性、计算机处理能力的适应性及不同属性的义项义频的差距等方面,都存在着相当大的差别。从而,可以对基于传统词典义项上形成的机用义库提出改造与完善的若干原则与方法。

Computer word sense tagging set up a new requirement fot traditional dictionary.Some researchers have long been aware of this issue,Yu Shi-Wen、Wu Yun-Fang proposed to use syntactic function as the main basis.Based on previous studies,the article first introduces the method for word sense tagging and seven resources.Word Sense Base for Word Sense Tagging of Modern Chinese（WSB） is in a key position.Then it proposes the differences on semantic analysis between traditional dictionary and machine dictionary,from the purpose of word sense tagging,coverage,semantic relationships,operability,adaptability to computer processing and frequency gap between sense of different part of speech.At last it discusses the principles and methods of building new word sense base for a higher degree general,written,large-scale real corpus of modern Chinese.It should be suitable for computer processing.