Topic segmentation with an aspect hidden Markov model(14)

We present a novel probabilistic method for partially unsupervised topic segmentation on unstructured text. Previous approaches to this problem utilize the hidden Markov model framework (HMM). The HMM treats a document as mutually independent sets of words

11

Figure7:Windowwidthvs.CoAPfortheHMMandAHMMintheNYTcorpusschemedescribedinsection4.2,wordsinthebeginningofthewindowareweightedmoreheavilythanwordstowardstheendofthewindow.Therefore,asthewindowsizeincreases,morewordsmakelessimpactontheobservationdistributionandthesegmenterdoesnotperformaswell.

TheHMMdoeswellonlargewindowssinceallwordsarecountedequally.How-ever,thisincreaseinperformanceisattheexpenseoflowsegmentationgranularity.WhiletheHMMperformsbetterthantheAHMMforlargewindows,itneverattainstheperformanceoftheAHMMinsmallwindows.Typically,theAHMMreachespeakperformanceatawindowsizeof10-15words.TheHMMbeginstoperformbetterthantheAHMMataround100words.

6Conclusionsandfuturework

Inthispaper,wehaveintroducedanewapproachtotextsegmentationusingauniqueprobabilisticmodelthatcombinesanaspectmodelwithanHMM.Thisisauni edframeworkwithinwhichwelearnbothdocumentclustersfortrainingandobservationprobabilitiesfornewsegmentations.TheAHMMdoeswellwithsmallwindowsofwordsallowingforamoreprecisesegmentationthanwiththeHMM.

Wehaveexperimentedwiththissystemonnoisytextsourcesproducedbyaspeechrecognitionsystem.Sinceourmodelispurelystatistical,wecansegmentthisoutputandaccuratelyhypothesizetopictransitionpoints.OurresultsontranscriptsproducedbytheSPEECHBOTsystemarequiteencouraging.

Futureworkinthisareahasseveraldirections.First,wewouldliketoincorporatesegmentationintotheSPEECHBOTIRframeworkinaprincipledwayandmeasureitssuccess.Second,wewouldliketousethetopiclabelstocategorizethecorpusofsegmentsandfurtherimproveaudiobrowsingandretrieval.Finally,wewouldliketo