一种基于语义词典的局部查询扩展方法-论文.pdf

《一种基于语义词典的局部查询扩展方法-论文.pdf》由会员上传分享，免费在线阅读，更多相关内容在应用文档-天天文库。

1、南京大学学报(自然科学)第50卷第4期Vo1．50，No．4JOURNALOFNAN『jINGUNIVERSITY2014年7月July，2014(NATURALSCIENCES)DoI：10．13232／j．cnki．jnju．2014．04．017一种基于语义词典的局部查询扩展方法吴秦，白玉昭，梁久祯(江南大学物联网工程学院，无锡，214122)摘要：针对基于关键词匹配的搜索引擎存在的问题，提出一种基于语义词典的局部查询扩展方法，首先利用共现分析法和语义相似度选取扩展词，再对原始查询词和扩展词加权，最后计算文档相似度从而获得排序后的

2、扩展查询结果．该方法克服了其它局部扩展方法将大量无关词加入查询的问题．实验表明，该方法有效地提高了查询结果的查准率．关键词：查询扩展，语义词典，共现分析，语义相似度AlocalqueryexpansionmethodbasedonsemanticdictionaryWuQin，BaiYuzhao，LiangJiuzhen(SchoolofInternetofThingsEngineering，JiangnanUniversity，Wuxi，214122，China)Abstract：Mosttraditionalsearchengine

3、modelsarebasedonkeywordmatching．Duetothelargenumberofsynonymsandpolysemouswords，thequeryresultsobtainedbytraditionalsearchengineshaveabigprobabilitytobedifferentfromwhattheuserexpected，especiallywhenthelengthofquerywordsisshort．Toovereomethisproblem，thispaperproposesanew

4、querymethodbasedonlocalqueryexpansiontechnologyandsemanticdictionary．Firstly，initialdocumentsetisobtainedbyquerywithoriginalkeywords．Andthedocumentsmostrelatedtotheoriginalkeywordsareselectedasextended—keyword-selectiondocuments．ByCO—OCCurrenceanalysis．WOrdswithlargeweig

5、htsareselectedasextendedkeywordcandidatesfromtheextended—keyword—seleetiondocuments．TongyiciCilin(ExtendedEdition)isusedasthesemanticdictionaryinthispaper．AccordingtothecharacteristicoftheencodingstyleofTongyiciCilin(ExtendedEdition)，anewmeasurementofwordsimnarityisdefin

6、ed。Anditisappliedtoselectextendedkeywordsfromtheextendedkeywordcandidates．Theoriginalkeywordsandtheextendedkeywordsareusedasthefinalquerywords．Togetbetterretrievalresults，eachwordinthefina1qefywordsetisassignedaweightbasedonitsimportanceinthequeryanditssimilaritytotheori

7、gina1keyWOrd．Thesimilaritiesbetweenthesetoffinalquerywordsandtheinitialdocumentsarecalculatedbasedontheweightsofwordsinthefinalquerywordset．Andthefinalretrievalresultsaresortedaccordingtothesimilariftesbetweenthesetoffina1querywordsandtheinitialdocuments．Comparingwithoth

8、erlocalqueryexpansionmethods，theproposedmethodavoidsaddingunrelatedwordstothequery．Totesttheeffectivene