【作者】 曹晶；

【导师】 孙铁利；

【作者基本信息】 东北师范大学 ， 计算机软件与理论， 2006， 硕士

【摘要】 随着Internet的飞速发展,Internet的应用深入到各行各业,Internet上的信息变得繁杂,信息的查找与获取也就变得越来越困难,如何从这个浩瀚的信息资源库中获取用户需要的信息,是人们面临的一个重要问题。传统的信息检索服务已不能满足用户的检索需求,因此智能信息检索成为重要的研究课题。概念检索是实现智能信息检索的重要方式之一。本文研究了搜索引擎特别是中文搜索引擎的相关现状,以及下一代智能搜索引擎即基于概念的检索系统的功能特点,并利用《同义词词林》作为语义体系实现了基于《同义词词林》的同义词扩展检索。并对同义词识别的算法进行了深入的研究,利用基于语义体系的相似度算法实现智能搜索引擎中的同义概念的识别,和相关词语的联想搜索。关键词检索是大多数搜索引擎的主要检索手段,然而多数搜索引擎对此不加控制。从情报语言学研究的成果以及网络信息检索的现状来看,自然语言不加以控制,检索效率与效果是不能尽如人意的。应该说,要克服关键词检索系统的弊病,必须对关键词进行有效的控制,变字面检索为概念检索。本文在分析目前概念检索系统的基础上,提出了基于同义词(概念)挖掘的概念检索系统。同义词的自动发现和识别在基于概念的信息检索领域有着重要的研究意义和应用价值,本文对国内和国外同义词识别算法进行研究和分析的基础上,对基于语义体系的同义词识别算法,即基于《同义词词林》的同义词识别算法和基于《知网》的同义词识别算法进行了深入的研究,利用词汇间的语义相似度度量来进行同义词识别,挖掘出大量的复合词形的同义词。基于《同义词词林》的同义词识别算法的实验结果表明,该算法具有较高的识别率。同义词的自动发现和识别,可以用于扩充同义词词典,提高同义词扩展检索的检准率和检全率。本文还在基于《知网》的词汇语义相似度算法的基础上提出了词语相关度计算方法,从而实现了相关概念联想功能。更多还原

【Abstract】 Along with the quick development of the Internet, the information on the Internet increases quickly everyday. How to obtain the information that customers want from this enormous resources databases, is an important problem for all users. The search engines are the most valid means at present. The intellectualized information retrieval has become a major research topic because the traditional information retrieval can not meet the requirements of users. Therefore, the concept-based retrieval is becoming one of the most important ways of intellectualized information retrieval.Firstly, this thesis summarizes the correlation research of Chinese search engine, based on this introduction, this thesis analyses the intellectualized information retrieval ,namely,the characteristic of concept-based retrieval system. We use the synonyms dictionary as the semantic system to realize synonyms expanding retrieval. Secondly we make a deep research of the algorithms of synonyms recognition. Using recognition synonym based on semantic system can make computer system recognition synonyms automatically. At the same time, we can realize concept association retrieval with it.Keyword retrieval is the most popular method in search engine retrieval, but most search engines do not control vocabulary in information retrieval. This makes the retrieval efficiency not competent. The resolution of this problem is to use“concept retrieval”instead of“word-form retrieval”. This thesis analyses the existing concept-based retrieval systems, based on this introduction, and puts forward a concept retrieval system based on synonyms recognition.Automatic recognition of synonyms plays an important role in concept-based information retrieval. Based on the analyses of the synonyms recognition, we use the similarity degree among the words to recognize synonyms, and mining a lot of multiple synonyms. Experiment has proved that this algorithm has well efficiency of recognition. The synonyms can be used to expand synonyms dictionary. It can improve the efficiency of the retrieval. Finally, this thesis provides concept association based on the algorithm of synonyms recognition.更多还原