分本分类
KnowledgeSupervisedTextClassi cationwith
NoLabeledDocuments
CongleZhang1,2,Gui-RongXue1,2,andYongYu
2ApexLab,ShanghaiJiaotongUniversity,Shanghai,200240StateKeyLabofCAD&CG,ZhejiangUniversity,Hangzhou,310058
{zhangcongle,grxue,yyu}@apex.sjtu.edu.cn1
Abstract.Intraditionaltextclassi cationapproaches,thesemanticmeaningsoftheclassesaredescribedbythelabeleddocuments.Sincela-belingdocumentsisoftentimeconsumingandexpensive,itisapromisingideathataskinguserstoprovidesomekeywordstodepicttheclasses,in-steadoflabelinganydocuments.However,shortpiecesofkeywordsmaynotcontainenoughinformationandthereforemayleadtounreliableclas-si er.Fortunately,therearelargeamountofpublicdataeasilyavailableinwebdirectories,suchasODP,Wikipedia,etc.Weareinterestedinexploringtheenormouscrowdintelligencecontainedinsuchpublicdatatoenhancetextclassi cation.Inthispaper,weproposeanoveltextclassi cationframeworkcalled“KnowledgeSupervisedLearning”(KSL),whichutilizestheknowledgeinkeywordsandthecrowdintelligencetolearntheclassi erwithoutanylabeleddocuments.Wedesignatwo-stageriskminimization(TSRM)approachfortheKSLproblem.Itcanopti-mizetheexpectedpredictionriskandbuildthehighqualityclassi er.Empiricalresultsverifyourclaim:ouralgorithmcanachieveabove0.9onMicro-F1onaverage,whichismuchbetterthanbaselinesandevencomparableagainstSVMclassi ersupervisedbylabeleddocuments.1Introduction
Supervisedlearning[2,5]andsemi-supervisedlearning[18,19]algorithmsarewidelyusedintextclassi cation.Theyrequirehighqualitylabeleddocumentsforeachclasstobuildtheclassi er.Somestudies[13]proposeanotherstrat-egythatuserscanproviderepresentativekeywordstodepicttheclasses.Suchalternativeisverypromisingbecauselabelingdocumentsisoftenmuchmoreex-pensivethangivingsomekeywords.Forexample,wheninternetserviceproviderswanttoclassifywebpagesaccordingtousers’interests,itismucheasiertoaskthemissuesomewordstodescribetheirinterestthantocollectwebpagestheyarereallyinterestedin.
However,onecanhardlyexpectuserstoprovidelonganddetailedkeywords.Itmeansthattheinformationofthekeywordsarenotenoughtolearnareli-ableclassi er.Thegapbetweenthekeywordsandthecorrespondingsemanticmeaningshouldbe lledup.Wenoticethatmanywebdirectoriesgrowsenor-mousandeasilyavailable,e.g.Opendirectoryproject(ODP),Wikipedia,Ya-hoo!Directory.SuchpublicdatasetscontainhugeamountofarticlestogetherT.-B.HoandZ.-H.Zhou(Eds.):PRICAI2008,LNAI5351,pp.509–520,2008.cSpringer-VerlagBerlinHeidelberg2008
Neighbour Classification with Background Knowledge extended to Semi-Supervised ...neighbour approach has also been shown to improve text classification when ...
Text Classification_专业资料。? Assignment of text documents to predefined ...8th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining...
[28] explored learning from symbolic knowledge. Among supervised learning ...Inherently, text classification problems have high dimensionality with very ...
slides_Robust supervised classification with mixture models 隐藏>> Robust supervised classi?cation with mixture models Learning from data with uncertain labels...
Improving Multiclass Text Classi?cation with the Support Vector Machine Jason D. M. Rennie & Ryan Rifkin Arti?cial Intelligence Laboratory Massachusetts ...
Semi-Supervised Discriminative Classification with Application to Tumorous Tissues Segmenta Due to the large data size of 3D MR brain images and the blurry...
After a brief introduction of the background knowledge for text classification problem in Section 2, the algorithm of feature selection combining OR+ SVM-...
Supervised Mineral Classification with Semiautomatic Training and Validation Set Generation This paper addresses the problem of classifying minerals common in ...
A Supervised Clustering Method for Text Classification Abstract. This paper describes a supervised three-tier clustering method for classifying students ’ ...
Support vector machine active learning with applications to text classification...Introduction In many supervised learning tasks, labeling instances to create a...
热门文档