Meta-classifier approach to reliable text classification(18)

A problem with automatic classifiers is that there is no way to know if a particular classification is just a guess or a certain answer. Reliable classification is the task of predicting whether a certain instance is correctly classified or not, i.e., a cl

Chapter3

TextClassi cation

Inthepreviouschapterwedescribedtheclassi cationtasksoftheCBSandthedatapreparationprocess.Intherestofthethesiswefocusonreliabletextclassi cationfortheCBS.The rststepinthemeta-classi erapproachistogenerateabaseclassi er.Thischapterstudiesthebasetextclassi ersderivedontheCBSdata.Weanswerour rstresearchquestion,thatisconcernedwith ndingane cientandaccuratetextclassi er.Section3.1elaboratesontextrepresentationandsection3.2discussesclassi cationalgorithms.Insection

3.3experimentswithtextclassi ersontheCBSdataaredescribed.Finally,insection3.4wegiveasummaryofthischapter.

3.1TextRepresentation

Textdocumentstypicallyconsistofstringsofcharactersthatcannotbedi-rectlyinterpretedbyaclassi er.Eachdocumentneedstobetransformedintoacompactrepresentationofitscontentappropriateforautomaticclassi cation.Thevectorrepresentationisthemostfrequentlyuseddocumentrepresenta-tion[Mladeni´c,1999].Adocumentwillberepresentedasavectorofweightsdj=<w1j,...,wmj>,whereMisthesetoffeaturesthatoccuratleastonceinatleastonedocumentinthedatasetandtheweights,0≤wkj≤1representhowmuchfeaturefkcontributestothesemanticsofdocumentdj.Di erentwaystode nefeaturesandtocomputefeatureweightsleadtodi erenttextrepresen-tationapproaches[Sebastiani,2002].Fourpossibilitiestode nefeaturesareasfollows.

rmationretrievalresearchsuggeststhatwordsworkwellasfeaturesandthattheirorderinginarecordisofminorimportanceformanytasks[Lewis,1992].Whentheweightsofthefeaturesarebinary,identifyingfeatureswithwordsiscalledthebag-of-wordsapproach.

ingphrasesasfeatureshasproventobenotverysuccessful.Di erentwaysofde ningphraseswereinvestigatedbyLewis[1992].

Wordn-grams.Afeatureisde nedforeachwordsequencecontaininguptonconsecutivewords.Thisrepresentationworksbestifthesequenceofwordsisimportantforthecontextandthedocumentsareshort[Mladeni´c,1998].

12