《LITERATURE-REVIEW文献综述.ppt》由会员分享，可在线阅读，更多相关《LITERATURE-REVIEW文献综述.ppt（47页珍藏版）》请在金锄头文库上搜索。
1、LITERATUREREVIEWDr.LiliAnnSMM4999-LiteratureReview2WhatisLR?NAliteraturereviewdiscussespublishedinformationinaparticularsubjectarea,andsometimesinformationinaparticularsubjectareawithinacertaintimeperiod.NAliteraturereviewcanbejustasimplesummaryofthesources,butitusuallyhasanorganizationalpatternandc
2、ombinesbothsummaryandsynthesis.Dr.LiliAnnSMM4999-LiteratureReview3WhatisLR?NAsummaryisarecapoftheimportantinformationofthesource,butasynthesisisare-organization,orareshuffling,ofthatinformation.NItmightgiveanewinterpretationofoldmaterialorcombinenewwitholdinterpretations.NOritmighttracetheintellectu
3、alprogressionofthefield,includingmajordebates.NAnddependingonthesituation,theliteraturereviewmayevaluatethesourcesandadvisethereaderonthemostpertinentorrelevant.Dr.LiliAnnSMM4999-LiteratureReview4WhatisLR?NTheformatofareviewofliteraturemayvaryfromdisciplinetodisciplineandfromassignmenttoassignment.N
4、Areviewmaybeaself-containedunit-anendinitself-oraprefacetoandrationaleforengaginginprimaryresearch.Areviewisarequiredpartofgrantandresearchproposalsandoftenachapterinthesesanddissertations.NGenerally,thepurposeofareviewistoanalyzecriticallyasegmentofapublishedbodyofknowledgethroughsummary,classifica
5、tion,andcomparisonofpriorresearchstudies,reviewsofliterature,andtheoreticalarticles.Dr.LiliAnnSMM4999-LiteratureReview5WhatisLR?NAliteraturereviewistheeffectiveevaluationofselecteddocumentsonaresearchtopic.NAreviewmayformanessentialpartoftheresearchprocessormayconstitutearesearchprojectinitself.NInt
6、hecontextofaresearchpaperorthesistheliteraturereviewisacriticalsynthesisofpreviousresearch.NTheevaluationoftheliteratureleadslogicallytotheresearchquestion.Dr.LiliAnnSMM4999-LiteratureReview6WhatisLR?Dr.LiliAnnSMM4999-LiteratureReview7WhywriteLR?NLiteraturereviewsprovideyouwithahandyguidetoaparticul
7、artopic.Ifyouhavelimitedtimetoconductresearch,literaturereviewscangiveyouanovervieworactasasteppingstone.NLiteraturereviewsalsoprovideasolidbackgroundforaresearchpapersinvestigation.Comprehensiveknowledgeoftheliteratureofthefieldisessentialtomostresearchpapers.Dr.LiliAnnSMM4999-LiteratureReview8Whyw
8、riteLR?NForprofessionals,theyareusefulreportsthatkeepthemuptodatewithwhatiscurrentinthefield.NForscholars,thedepthandbreadthoftheliteraturereviewemphasizesthecredibilityofthewriterinhisorherfieldDr.LiliAnnSMM4999-LiteratureReview9WhywriteLR?NThepurposeofaliteraturereviewisforyoutotakeacriticallookat
9、theliterature(factsandviews)thatalreadyexistsintheareayouareresearching.NAliteraturereviewisnotashoppinglistofeverythingthatexists,butacriticalanalysisthatshowsanevaluationoftheexistingliteratureandarelationshipbetweenthedifferentworks.NItdemonstratestherelevanceoftheresearch.Dr.LiliAnnSMM4999-Liter
10、atureReview10WhywriteLR?NLiteraturecanincludebooks,journalarticles,internet(electronicjournals),newspapers,magazines,thesesanddissertations,conferenceproceedings,reports,anddocumentaries.NLiteraturereviewsarewrittenoccasionallyinthehumanities,butmostlyinthesciencesandsocialsciences;inexperimentandla
11、breports,theyconstituteasectionofthepaper.NSometimesaliteraturereviewiswrittenasapaperinitself.Dr.LiliAnnSMM4999-LiteratureReview11WhywriteLR?NInthecontextofaresearchpaperonathesis,theliteraturereviewprovidesabackgroundtothestudybeingproposed.NThebackgroundmayconsideroneormoreofthefollowingaspectsde
12、pendingontheresearchquestionbeingposed:NTheoreticalbackgroundpast,presentorfutureNClinicalpracticepreviousorcontemporaryNMethodologyand/orresearchmethodsNPreviousfindingsNRationaleand/orrelevanceofthecurrentstudyDr.LiliAnnSMM4999-LiteratureReview12WhywriteLR?NInabroadercontextHart(1998)liststhefollo
13、wingpurposesofareview:NDistinguishingwhathasbeendonefromwhatneedstobedone;NDiscoveringimportantvariablesrelevanttothetopic;NSynthesisingandgaininganewperspective;NIdentifyingrelationshipsbetweenideasandpractice;NEstablishingthecontextofthetopicorproblem;Dr.LiliAnnSMM4999-LiteratureReview13WhywriteLR
14、?NRationalisingthesignificanceoftheproblem;NEnhancingandacquiringthesubjectvocabulary;NUnderstandingthestructureofthesubject;NRelatingideasandtheorytoapplications;NIdentifyingmethodologiesandtechniquesthathavebeenused;NPlacingtheresearchinahistoricalcontexttoshowfamiliaritywithstate-of-the-artdevelo
15、pments.Dr.LiliAnnSMM4999-LiteratureReview14WhywriteLR?NItspurposeisto:NPlaceeachworkinthecontextofitscontributiontotheunderstandingofthesubjectunderreviewNDescribetherelationshipofeachworktotheothersunderconsiderationNIdentifynewwaystointerpret,andshedlightonanygapsin,previousresearchNResolveconflic
16、tsamongstseeminglycontradictorypreviousstudiesNIdentifyareasofpriorscholarshiptopreventduplicationofeffortNPointthewayforwardforfurtherresearchNPlaceonesoriginalwork(inthecaseofthesesordissertations)inthecontextofexistingliteratureDr.LiliAnnSMM4999-LiteratureReview15WhatshouldIdobeforewritingthelite
17、raturereview?NClarifyNIfyourassignmentisnotveryspecific,seekclarificationfromyoursupervisor/lecturer:NRoughlyhowmanysourcesshouldyouinclude?NWhattypesofsources(books,journalarticles,websites)?NShouldyousummarize,synthesize,orcritiqueyoursourcesbydiscussingacommonthemeorissue?NShouldyouevaluateyourso
18、urces?NShouldyouprovidesubheadingsandotherbackgroundinformation,suchasdefinitionsand/orahistory?Dr.LiliAnnSMM4999-LiteratureReview16WhatshouldIdobeforewritingtheliteraturereview?NFind modelsNLookforotherliteraturereviewsinyourareaofinterestorinthedisciplineandreadthemtogetasenseofthetypesofthemesyou
19、mightwanttolookforinyourownresearchorwaystoorganizeyourfinalreview.YoucansimplyputthewordreviewinyoursearchenginealongwithyourothertopictermstofindarticlesofthistypeontheInternetorinanelectronicdatabase.Thebibliographyorreferencesectionofsourcesyouvealreadyreadarealsoexcellententrypointsintoyourownr
20、esearch.Dr.LiliAnnSMM4999-LiteratureReview17WhatshouldIdobeforewritingtheliteraturereview?NNarrow your topicNTherearehundredsoreventhousandsofarticlesandbooksonmostareasofstudy.Thenarroweryourtopic,theeasieritwillbetolimitthenumberofsourcesyouneedtoreadinordertogetagoodsurveyofthematerial.Yourinstru
21、ctorwillprobablynotexpectyoutoreadeverythingthatsoutthereonthetopic,butyoullmakeyourjobeasierifyoufirstlimityourscope.Dr.LiliAnnSMM4999-LiteratureReview18WhatshouldIdobeforewritingtheliteraturereview?NConsider whether your sources are currentNSomedisciplinesrequirethatyouuseinformationthatisascurren
22、taspossible.Inthesciences,forinstance,treatmentsformedicalproblemsareconstantlychangingaccordingtothelateststudies.Informationeventwoyearsoldcouldbeobsolete.Dr.LiliAnnSMM4999-LiteratureReview19WhatshouldIdobeforewritingtheliteraturereview?NHowever,ifyouarewritingareviewinthehumanities,history,orsoci
23、alsciences,asurveyofthehistoryoftheliteraturemaybewhatisneeded,becausewhatisimportantishowperspectiveshavechangedthroughtheyearsorwithinacertaintimeperiod.NTrysortingthroughsomeothercurrentbibliographiesorliteraturereviewsinthefieldtogetasenseofwhatyourdisciplineexpects.NYoucanalsousethismethodtocon
24、siderwhatishotandwhatisnot.Dr.LiliAnnSMM4999-LiteratureReview20WhatshouldIdobeforewritingtheliteraturereview?NFind a focusNAliteraturereview,likeatermpaper,isusuallyorganizedaroundideas,notthesourcesthemselvesasanannotatedbibliographywouldbeorganized.Thismeansthatyouwillnotjustsimplylistyoursourcesa
25、ndgointodetailabouteachoneofthem,oneatatime.NNo.Dr.LiliAnnSMM4999-LiteratureReview21WhatshouldIdobeforewritingtheliteraturereview?NAsyoureadwidelybutselectivelyinyourtopicarea,considerinsteadwhatthemesorissuesconnectyoursourcestogether.NDotheypresentoneordifferentsolutions?NIsthereanaspectofthefield
26、thatismissing?NHowwelldotheypresentthematerialanddotheyportrayitaccordingtoanappropriatetheory?NDotheyrevealatrendinthefield?NAragingdebate?NPickoneofthesethemestofocustheorganizationofyourreview.Dr.LiliAnnSMM4999-LiteratureReview22WhatshouldIdobeforewritingtheliteraturereview?NConstruct a working t
27、hesis statementNThenusethefocusyouvefoundtoconstructathesisstatement.Yes!Literaturereviewshavethesisstatementsaswell!However,yourthesisstatementwillnotnecessarilyargueforapositionoranopinion;ratheritwillargueforaparticularperspectiveonthematerial.Dr.LiliAnnSMM4999-LiteratureReview23WhatshouldIdobefo
28、rewritingtheliteraturereview?NSomesamplethesisstatementsforliteraturereviewsareasfollows:NThe current trend in treatment for congestive heart failure combines surgery and medicine.NMore and more cultural studies scholars are accepting popular media as a subject worthy of academic consideration.Dr.Li
29、liAnnSMM4999-LiteratureReview24WhatshouldIdobeforewritingtheliteraturereview?NConsider organizationNYouvegotafocus,andyouvenarroweditdowntoathesisstatement.NNowwhatisthemosteffectivewayofpresentingtheinformation?NWhatarethemostimportanttopics,subtopics,etc.,thatyourreviewneedstoinclude?NAndinwhatord
30、ershouldyoupresentthem?Dr.LiliAnnSMM4999-LiteratureReview25WhatshouldIdobeforewritingtheliteraturereview?NDevelopanorganizationforyourreviewatbothaglobalandlocallevel:NFirst, cover the basic categoriesNJustlikemostacademicpapers,literaturereviewsalsomustcontainatleastthreebasicelements:anintroductio
31、norbackgroundinformationsection;thebodyofthereviewcontainingthediscussionofsources;and,finally,aconclusionand/orrecommendationssectiontoendthepaper.Dr.LiliAnnSMM4999-LiteratureReview26WhatshouldIdobeforewritingtheliteraturereview?NIntroduction:Givesaquickideaofthetopicoftheliteraturereview,suchasthe
32、centralthemeororganizationalpattern.NBody:Containsyourdiscussionofsourcesandisorganizedeitherchronologically,thematically,ormethodologically(seebelowformoreinformationoneach).NConclusions/Recommendations:Discusswhatyouhavedrawnfromreviewingliteraturesofar.Wheremightthediscussionproceed?Dr.LiliAnnSMM
33、4999-LiteratureReview27WhatshouldIdobeforewritingtheliteraturereview?NThe introductionshouldprovidethereaderwiththescaleandstructureofyourreview.Itservesasakindofmap.NThe bodyofthereviewdependsonhowyouhaveorganisedyourkeypoints.Literaturereviewsatpostgraduatelevelshouldbeevaluativeandnotmerelydescri
34、ptive.Forexamplepossiblereasonsforsimilaritiesordifferencesbetweenstudiesareconsideredratherthanamereidentificationofthem.NThe conclusionofthereviewneedstosumupthemainfindingsofyourresearchintotheliterature.Thefindingscanberelatedtotheaimsofthestudyyouareproposingtodo.Thereaderisthusprovidedwithacoh
35、erentbackgroundtothecurrentstudy.Dr.LiliAnnSMM4999-LiteratureReview28WhatshouldIdobeforewritingtheliteraturereview?NOrganizing the bodyNOnceyouhavethebasiccategoriesinplace,thenyoumustconsiderhowyouwillpresentthesourcesthemselveswithinthebodyofyourpaper.Createanorganizationalmethodtofocusthissection
36、evenfurther.NTohelpyoucomeupwithanoverallorganizationalframeworkforyourreview,considerthesixtypicalwaysoforganizingthesourcesintoareview:NChronologicalNBy publicationNBy trendNThematicNMethodologicalNQuestions for Further ResearchDr.LiliAnnSMM4999-LiteratureReview29WhatshouldIdobeforewritingtheliter
37、aturereview?NSimilartoprimaryresearch,developmentoftheliteraturereviewrequiresfourstages:NProblemformulationwhichtopicorfieldisbeingexaminedandwhatareitscomponentissues?NLiteraturesearchfindingmaterialsrelevanttothesubjectbeingexploredNDataevaluationdeterminingwhichliteraturemakesasignificantcontrib
38、utiontotheunderstandingofthetopicNAnalysisandinterpretationdiscussingthefindingsandconclusionsofpertinentliteratureDr.LiliAnnSMM4999-LiteratureReview30Whatshouldyouwrite?NtheacceptedfactsintheareaNthepopularopinionNthemainvariablesNtherelationshipbetweenconceptsandvariablesNshortcomingsintheexisting
39、findingsNlimitationsinthemethodsusedintheexistingfindingsNtherelevanceofyourresearchNsuggestionsforfurtherresearchinthearea.Dr.LiliAnnSMM4999-LiteratureReview31Whatshouldyouwrite?NLiteraturereviewsshouldcomprisethefollowingelements:NAnoverviewofthesubject,issueortheoryunderconsideration,alongwiththe
40、objectivesoftheliteraturereviewNDivisionofworksunderreviewintocategories(e.g.thoseinsupportofaparticularposition,thoseagainst,andthoseofferingalternativethesesentirely)NExplanationofhoweachworkissimilartoandhowitvariesfromtheothersNConclusionsastowhichpiecesarebestconsideredintheirargument,aremostco
41、nvincingoftheiropinions,andmakethegreatestcontributiontotheunderstandinganddevelopmentoftheirareaofresearchDr.LiliAnnSMM4999-LiteratureReview32Whatshouldyouwrite?NInassessingeachpiece,considerationshouldbegivento:NProvenanceWhataretheauthorscredentials?Aretheauthorsargumentssupportedbyevidence(e.g.p
42、rimaryhistoricalmaterial,casestudies,narratives,statistics,recentscientificfindings)?NObjectivityIstheauthorsperspectiveeven-handedorprejudicial?Iscontrarydataconsideredoriscertainpertinentinformationignoredtoprovetheauthorspoint?NPersuasivenessWhichoftheauthorsthesesaremost/leastconvincing?NValueAr
43、etheauthorsargumentsandconclusionsconvincing?Doestheworkultimatelycontributeinanysignificantwaytoanunderstandingofthesubject?Dr.LiliAnnSMM4999-LiteratureReview33Whatshouldyouwrite?NLayoutNMakeyourliteraturereviewhaveanacademicandprofessionalappearance.Herearesomepointstomakethelookofyourreportappeal
44、ingtothereaderNWhite space:leavespacebetweensections,especiallyfromtheabstract.Thisgivesanunclutteredeffect.NHeadings/sub-headings:thesehelptoseparateideas.NText boxes:youcanusetheseforquotationsorparaphrasingtoseparatethemfromtherestofyourtext.Itisalsopleasingtotheeye.Dr.LiliAnnSMM4999-LiteratureRe
45、view34Whatshouldyouwrite?NGraphics:centreyourgraphics,suchasdiagramsortables,tohavespacearoundthem.Trynottoburygraphicsinyourtext.NPagination:youcannumberpagesorsectionsorboth,buttheimportantthingtodoistobeconsistent.Thecoverpagenormallyisnotnumbered.Thecontentpageandabstractpageusuallyhaveaseparate
46、numberingsystemtothebodyofyourliteraturereview.Dr.LiliAnnSMM4999-LiteratureReview35Whatshouldyouwrite?NLanguage focusNCreateabalancebetweendirectquotation(citation)andparaphrasing.Avoidtoomuchdirectquoting.Theverbtensechosendependsonyouremphasis:NWhenyouarecitingaspecificauthorsfindings,usethepastte
47、nse:(found,demonstrated);NWhenyouarewritingaboutanacceptedfact,usethepresenttense:(demonstrates,finds);andNWhenyouarecitingseveralauthorsormakingageneralstatement,usethepresentperfecttense:(haveshown,havefound,littleresearchhasbeendone).Dr.LiliAnnSMM4999-LiteratureReview36Whatshouldyouwrite?NFinal c
48、hecklistNHaveIfulfilledthepurposeoftheliteraturereview?NIsitwrittenatalevelappropriatetoitsaudience?NAreitsfactscorrect?NIsalltheinformationincludedrelevant?NArethelayoutandpresentationeasyontheeye?NIsthelanguageclear,conciseandacademic?NDoestheabstractsummarisetheentirereview?NDoestheintroductionad
49、equatelyintroducethetopic?NIsthebodyorganisedlogically?NDoestheconclusioninterpret,analyseandevaluate?NAretherecommendationsreasonable?NDoesthetableofcontentscorrespondwiththeactualcontents?Arepagenumberscorrect?NHaveIacknowledgedallsourcesofinformationthroughcorrectreferencing?NHaveIcheckedspelling
50、,grammarandpunctuation?NHaveIcarefullyproof-readthefinaldraft?Dr.LiliAnnSMM4999-LiteratureReview37Howtoreview?NThewholeprocessofreviewingincludes:a.Searchingforliteratureb.Sortingandprioritisingtheretrievedliteraturec.Analyticalreadingofpapersd.Evaluativereadingofpaperse.Comparisonacrossstudiesf.Org
51、anisingthecontentg.WritingthereviewDr.LiliAnnSMM4999-LiteratureReview38Howtoreview?NComparison across studiesNTheaimistoextractkeypointsbycomparingandcontrastingACROSSstudies,insteadofreadingonepaperafteranother.NKeypointsforareviewmayconcernareasofsimilaritiesand/ordifferencesin:NResearchaim(s)orhy
52、pothesesNResearchdesignandsamplingNInstrumentsandproceduresusedNHowdatawereanalysedNResultsorfindingsNInterpretationsDr.LiliAnnSMM4999-LiteratureReview39Howtoreview?NFindsimilaritiesanddifferencesbetweenstudiesatdifferent levels,e.g.:-philosophy-epistemology-morality-methodology-methods-typesofdata-
53、dataanalysis-interpretationDr.LiliAnnSMM4999-LiteratureReview40Howtoreview?NSetoutyourthinkingonpaperthroughmapsandtrees.Dr.LiliAnnSMM4999-LiteratureReview41Howtoreview?NTipsonwritingDr.LiliAnnSMM4999-LiteratureReview42Howtoreview?oPitfalls - Vaguenessduetotoomuchorinappropriategeneralisations-Limit
54、edrange-Insufficientinformation-Irrelevantmaterial-Omissionofcontrastingview-OmissionofrecentworkDr.LiliAnnSMM4999-LiteratureReview43ExampleEarlyworkshaveaddressedsomeoftheproblemsandissuesdiscussedinvideoretrieval.Researchershavedevelopedideasandtoolsforsupportingvideoediting,forexamplein8.Theyhave
55、definedaseamlessvideoeditinginthegradientdomain.Thespatio-temporalgradientfieldsoftargetvideosaremodifiedormixedtogenerateanewgradientfield,whichisusuallynotintegrateable.Theyhavealsodescribedhowsemanticinformationaboutvideocanbestructuredandusedforcontent-basedaccess.Fromageneralvideoarchivepointof
56、view,theproblemwiththistoolisthelackofsupportformanagingvideodocumentstructures.AdigitalvideoarchiveservingdifferentcategoriesofusersshouldofferamorestructuredwayofdescribingvideocontentsDr.LiliAnnSMM4999-LiteratureReview44ExampleHiddenMarkovModels(HMMs)arestatisticaltoolsthathavebeenusedsuccessfull
57、yinmodellingdifficulttaskssuchasspeechrecognition15orbiologicalsequenceanalysis16.Inspiredbyasimilarspeechapplication,HiddenMarkovmodel(HMM)hasalsobeenappliedtoactivityrecognition.ThefirstapproachforthehumanmovementsbasedonHMMswasdescribedin13.Itdistinguishedbetweensixdifferenttennisstrokes.Thissyst
58、emdividedtheimageintomeshesandcountedthenumberofpixelsrepresentingthepersonforeachmesh.Thenumberswerecomposedtoafeaturevectorthatwasconvertedintoadiscretelabelbyavectorquantizer.ThelabelswereclassifiedbasedondiscreteHMMs.In8,anHMMisusedasarepresentationofsimpleactionswhicharerecognizedbycomputingthe
59、probabilitythatthemodelproducesthevisualobservationsequence.In14layeredHMMswereproposedtomodelsinglepersonofficeactivitiesatvarioustimegranularitiesDr.LiliAnnSMM4999-LiteratureReview45ExampleMostoftheexistingworkreliesonusingonlyasinglesourceofinformation(example,eitheraudioorvisualtrackdataalone).I
60、n4,theaveragevideoshotactivityandthedurationareusedasfeaturesforthecategorizationofmoviesaccordingtotheactions.Anactionscenewascharacterizedbytemporallylocalizedpropertiesofvideoshotswhichhavelittleornorecurringsimilarvisualcontents5.Althoughthesevisualcharactersareundoubtedlygoodindicatorsofrapidly
61、evolvingactioncontents,theyarenotenoughtodeterminethedesiredaction.Ontheotherhand,audio-basedactiondetectionwasindependentlyperformedonthesoundtrackin6.However,thisaudioalonemethodmayleadtomanypotentialfalsedetectedcasesbecausemanysoundsoftenmixdifferentnoisesandothersimilarbackgroundsound.Dr.LiliAn
62、nSMM4999-LiteratureReview46ExampleVideoPre-processingClassificationMotionGenreObjectEventShotColour Feature ExtractionEdge Feature ExtractionObjectCameraLengthDurationColour HistogramEdge DetectionEdge ClusteringPanZoomTranslationEdge CountRule-based classifiersRecognition resultFIGURE2.3SchematicdiagramforvideoclassificationSource:L.N.Abdullahetal.2005.Dr.LiliAnnSMM4999-LiteratureReview47Example
