资源描述：
《数据标准化与随机游走下的语义关系相似度计算》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。
1、复旦大学硕:1:学位论义AbstractAbstractItisanimportantresearchtopictomeasurethesemanticrelationalsimilaritybetweentwowordpairs,whichhasgreatvaluesinapplicationssuchassemanticsearch,informationextraction,analogydetectionandontologyconstruction.Theexistingmethodstomeasuringrelationalsimilaritycanb
2、eroughlydividedintotwocategories:theonebasedonsemanticresources(suchasWordNet),andtheotherbasedonlarge-scaletext(whichisnormallystatistical).Thestatisticalmethodsbasedonlarge-scaletextcorpusfirstextractrelatedlexical/syntacticpatternsaccordingtothecontextswherethewordpairco-occur,then
3、calculatethefrequenciesofthoseextractedpatternsbetweenthewordpair，andfinallyderiveouttherelationalsimilaritybetweendifferentwordpairs.Thiskindofmethodsoftensuffersfromtheproblemofdatasparsity.Thispaperhasanalyzedandsummarizedtheprocessingflowsandkeytechniquesusedbyclassicalmethods,and
4、hasdonethefollowingwork:Firstly,duetothefactthatdatastandardizationtechniqueplaysanimportantroleinrelationalsimilaritymeasurement,thispaperstudiedtheapplicationofthreedifferentdatastandardizationmethods(theintervalstandardization,thezScorestandardization,andtheentropy-weightedstandard
5、ization),andevaluatedtheireffectivenessontheENTdatasetandtheSATanalogiesquestions.Secondly,sincethestatisticalmethodsbasedonlarge-scaletextcorpusoftensufferfromtheproblemofdatasparsity,weemployedtherandomwalkalgorithmtosolvetheproblem,whichwasalsotestedandanalyzedbyexperiments.Finally
6、,thispaperproposedaprocessingflowtointegratetheautomatictermrecognitionandtherelationalsimilaritymeasurement,whichprovidesamechanismforautomaticacquisitionofwordpairs.Keywords:Semanticrelationship,Semanticsimilarity,Datastandardization,Randomwalk,Automatictermrecognition.Classificatio
7、nCode:TP391.1复旦大学硕：I：研究生学位论文第一章绪论第一章绪论如今互联网信息飞速增长，面对如此海量的信息，如何自动获取并蹄选更贴近人们所需的信息是知识发现领域的重要课题，而在词对语义关系相似度方面的研究更是重中之重，它在关系提取、信息检索、类比检测、查询扩展、问答系统和本体网络构建等领域都有着广泛的应用。近年来，针对词对间语义关系的研究取得了巨大的突破。本文主要针对英语词对之间蕴含的语义关系的相似度度量的问题进行探索，同时将已有的算法理论和词对间蕴含的语义关系相似度度量问题的一些特质相结合提出了解决该问题的新方法。1.1研究背景词是词对的基本单元，
8、而词对则是
