从Web网页上获取一价事件常识的方法  

Gainting Method Univalence Thing from Web Page

作　　者：杨帆[1] 孙强[2] 

机构地区：[1]江苏科技大学 [2]91206部队

出　　处：《科学技术与工程》2010年第25期6300-6304,共5页Science Technology and Engineering

摘　　要：一价事件是以一价动词为核心构成的事件。为了提高查询的智能性和准确性,尝试用一价事件设计描述了从《实习词表》中挑选一价动词,根据《一价动词表》设计Google查询项,根据Google查询项从Web网页上获取大规模的语料。用特征提取方法从Web语料中抽取事件上下文中的相关词,根据相关词评价Google查询项的优劣,并进行调整。得到与一价动词相关的因果逻辑,丰富查询项,从而提高查询精度。Univalence thing is composed of univalence verb.For improving the searching intelligence and accuracy,using univalence thing to describe picking univalence verb form ＂practical verb table is acceptd to＂ ,designing google searching item according to ＂univalence verb table＂ and gaining extensive language material from Web page.Use feature extraction method to extract interrelated verb from context.Evaluating and adjusting the Google searching item according to related word. Causal logic words can be gain which are related to univalence verb. The causal logic words will be used to enrich the searching item and proved the searching precision.

关 键 词：特征提取 文本 信息检索 分词算法 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]