ContextSemantic-basedNaiveBayesianAlgorithmforTextClassification

Abstract

Abstract: TheNaiveBayesclassifierisbasedontheassumptionthatthesamples’attributesareindependentoneanother.Asasimplebag-of-wordsmodel,itignorestheinfluenceofsynonymsincontexttoclassification.Thispaperproposestheconceptofsimilarwordandusesclustersofsimilarwordsinsteadofkeyworddictionaryintraining.First,word2vecistrainedtogetwordembedding.Second,thekeyworddictionaryisrepresentedbywordembeddingwhichisthenclusteredhierarchically,theclustersofsimiliarwordsarebuiltandexpanded.Theexperimentalresultsshowthattheabovemethodcanimprovetheaccuracyoftextclassification,andavoidtheinstabilityofclassificationeffectduetothedifferencesintrainingcorpus.

Key words: textcategorization, NaiveBayes, word2vec

CLC Number:

TP391

ZHENGKai-yu,ZHUCui. ContextSemantic-basedNaiveBayesianAlgorithmforTextClassification[J]. Computer and Modernization, 2018, 0(06): 1-.

References

［1］张雯,张化祥.属性加权的朴素贝叶斯集成分类器［J］.计算机工程与应用,2010,46(29):144-146.
［2］JiangJJ,ConrathDW.Semanticsimilaritybasedoncorpusstatisticsandlexicaltaxonomy［C］//Proceedingsof1997IEEEInternationalConferenceonResearchinComputationalLinguistics.1997：19-33.
［3］郭永辉.面向短文本分类的特征扩展方法［D］.哈尔滨：哈尔滨工业大学,2013.
［4］ZhangWeitai,XuWeiran,ChenGuang,etal.Afeatureextractionmethodbasedonwordembeddingforwordsimilaritycomputing［M］//NaturalLanguageProcessingandChineseComputing.Springer,2014:160-167.
［5］张东娜,周春光,刘彦斌,等.一种基于WordNet和CorpusStatistics的语义相似性计算方法［J］.吉林大学学报(理学版),2010,48(5):811-816.
［6］袁满,欧阳元新,熊璋,等.一种基于频繁词集的短文本特征扩展方法［J］.东南大学学报(自然科学版),2014,44(2):256-260.
［7］AgrawalR,ImielinskiT,SwamiA.Databasemining:Aperformanceperspective［J］.IEEETransactionsonKnowledgeandDataEngineering,1993,5(6):914-925.
［8］FriedmanN,GeigerD,GoldszmidtM.Bayesiannetworkclassifiers［J］.MachineLearning,1997,29(2-3):131-163.
［9］〖JP2〗MladenicD,GrobelnikM.FeatureselectionforunbalancedclassdistributionandNaiveBayes［C］//Proceedingsofthe16thInternationalConferenceonMachineLearning.1999:258-267
［10］YangYiming,PedersenJO.Acomparativestudyonfeatureselectionintextcategorization［C］//Proceedingsofthe14thInternationalConferenceonMachineLearning.1997:412-420.
［11］SaltonG,WongA,YangCS.Avectorspacemodelforautomaticindexing［J］.CommunicationsoftheACM,1975,18(11):613-620.
［12］MikolovT,SutskeverI,ChenKai,etal.Distributedrepresentationsofwordsandphrasesandtheircompositionality［C］//Proceedingsofthe26thInternationalConferenceonNeuralInformationProcessingSystems.2013,2:3111-3119
［13］MikolovT,YihWT,ZweigG.Linguisticregularitiesincontinuousspacewordrepresentations［C］//Proceedingsof2013ConferenceoftheNorthAmericanChapteroftheAssociationforComputationalLinguistics:HumanLanguageTechnologies.2013:746-751.
［14］杨婉霞,孙理和,黄永峰.结合语义与统计的特征降维短文本聚类［J］.计算机工程,2012,38(22):171-175.
［15］WolfL,HananiY,BarK,etal.Jointword2vecnetworksforbilingualsemanticrepresentations［J］.InternationalJournalofComputationalLinguisticsandApplications,2014,5(1):27-44.［16］MikolovT,ChenKai,CorradoG,etal.Efficientestimationofwordrepresentationsinvectorspace［J］.ComputerScience,2013（1）：28-36.
［17］JoulinA,GraveE,BojanowskiP,etal.Bagoftricksforefficienttextclassification［C］//Proceedingsofthe15thConferenceoftheEuropeanChapteroftheAssociationforComputationalLinguistics.2017:427-431.
［18］LillebergJ,ZhuYun,ZhangYanqing.Supportvectormachinesandword2vecfortextclassificationwithsemanticfeatures［C］//Proceedingsofthe14thIEEEInternationalConferenceonCognitiveInformatics&CognitiveComputing.2015:136-140.

[1]	CHEN Ying, SHENG Jia-gen. Weibo Tag Generation Algorithm Based on LDA and Word2vec [J]. Computer and Modernization, 2021, 0(12): 37-42.
[2]	JING Dong-sheng, XUE Jing-song, FENG Ren-jun. Spam Text Classification Method Based on Deep Q-network [J]. Computer and Modernization, 2020, 0(06): 89-.
[3]	NIU Xue-ying. CNN Text Classification Based on Topic Model Word Vectors [J]. Computer and Modernization, 2019, 0(10): 7-.
[4]	DONG Ao-gen, LIU Mao-fu, HUANG Ge-xin, SHU Qi-yun. Automatic Discovery of Relationship Between Examination Question and Knowledge Points Based on Vector Space Model [J]. Computer and Modernization, 2015, 0(10): 6-9.