




版權(quán)說明:本文檔由用戶提供并上傳,收益歸屬內(nèi)容提供方,若內(nèi)容存在侵權(quán),請進(jìn)行舉報(bào)或認(rèn)領(lǐng)
文檔簡介
FeatureSelectionforRanking
Cenfangming2007.10ABSTRACT
TherealityisthatmanyfeatureselectionmethodsusedinclassificationaredirectlyappliedtorankingWearguethatbecauseofthestrikingdifferencesbetweenrankingandclassification,itisbettertodevelopdifferentfeatureselectionmethodsforranking.
InthispaperweproposeanewfeatureselectionmethodDefinetherankingaccuracyintermsofaperformancemeasureoralossfunctionastheimportanceofthefeatureandthecorrelationbetweentherankingresultsoftwofeaturesasthesimilaritybetweenthem.Wealsodemonstratehowtosolvetheoptimizationprobleminanefficientway.Wehavetestedtheeffectivenessofourfeatureselectionmethodontwoinformationretrievaldatasetsandwithtworankingmodel1.INTRODUCTION
RankingSVMandRankNet
First,featureselectioncanhelpenhanceaccuracyinmanymachinelearningproblems,whichstronglyindicatesthatfeatureselectionisalsonecessaryforranking
Second,featureselectioncanalsohelpimprovetheefficiencyoftraining.Ininformationretrieval,especiallyinwebsearch,usuallythedatasizeisverylargeandthustrainingofrankingmodelsiscomputationallycostly.
Therehavebeennomethodsoffeatureselectiondedicatedlyproposedforranking.Mostofthemethodsusedinrankingweredevelopedforclassification.Featureselectionmethodsinclassificationfallintothreecategorie
Thefirstcategory,whichisnamedfilter,featureselectionisdefinedasapreprocessingstepandcanbeindependentfromlearning.thatinformationgain(IG)andchi-square(CHI)areamongthemosteffectivemethodsoffeatureselectionforclassification.
ThesecondcategoryreferredtoaswrapperutilizesthelearningsystemasablackboxtoscoresubsetsoffeaturesThethirdcategorycalledtheembeddedmethodperformsfeatureselectionwithintheprocessoftraining.Second,theevaluationmeasures(e.g.meanaverageprecision(MAP)andnormalizeddiscountedcumulativegain(NDCG))usedinrankingproblemsaredifferentfromthosemeasuresusedinclassification
1)inrankingusuallyprecisionismoreimportantthanrecallwhileinclassificationbothprecisionandrecallareimportant;
SeveralproblemsmayarisewhenapplyingthefeatureselectionmethodstorankingFirst:existingfeatureselectionmethodsforclassificationarenotsuitableforranking.Inranking,anumberoforderedcategoriesareused,representingtherankingrelationshipbetweeninstances,whileinclassificationthecategoriesare“flat”
2)
inrankingcorrectlyrankingthetop-ninstancesismorecriticalwhileinclassificationmakingacorrectclassificationdecisionisofequalsignificanceforallinstances.
Thefollowingarethepropertiesofthenovelmethodforfeatureselectioninranking1)Themethodmakesuseofrankinginformation,insteadofsimplyviewingtheranksasflatcategories.2)Inspiredbytheworkin[1][14][27],itconsidersthesimilaritiesbetweenfeatures,andtriestoavoidselectingredundantfeatures.3)Itmodelsfeatureselectionforrankingasamulti-objectiveoptimizationproblem.Thefinalobjectiveistofindasetoffeatureswithmaximumimportanceandminimumsimilarity.4)Itprovidesagreedysearchalgorithmtosolvetheoptimizationproblem.Thecorrespondingsolutionproducedisproventobeequivalenttotheoptimalsolutiontotheoriginalproblemundercertaincondition.2.FEATURESELECTIONMETHOD
2.1OverviewSupposethegoalistoselect(1≤t≤m)featuresfromtheentirefeatureset{1,2,…,}.Inourmethodwefirstdefinetheimportancescoreofeachfeature,anddefinethesimilaritybetweenanytwofeaturesand.Thenweemployanefficientalgorithmtomaximizethetotalimportancescoresandminimizethetotalsimilarityscoresofasetoffeatures.
2.2ImportanceoffeatureWefirstassignanimportancescoretoeachfeature.Specifically,weproposeusinganevaluationmeasurelikeMAPandNDCGoralossfunction(e.g.pair-wiserankingerrors[10][13])tocomputetheimportancescore.Intheformer,wefirstrankinstancesusingthefeature,evaluatetheperformanceintermsofthemeasure,andthentaketheevaluationresultastheimportancescore.Inthelatter,wealsorankinstancesusingthefeature,andthenviewascoreinverselyproportionaltothecorrespondinglossastheimportancescore
2.3SimilaritybetweenfeaturesInthiswork,wemeasurethesimilaritybetweenanytwofeaturesonthebasisoftheirrankingresults.Thatis,weregardeachfeatureasarankingmodel,andthesimilaritybetweentwofeaturesisrepresentedbythesimilaritybetweentherankingresultsthattheyproduce.wechooseKendall’s??asanexample.TheKendall’s??valueofqueryqforanytwofeatures????and????canbecalculatedasfollows,
2.4Optimizationformulation
wewanttoselectthosefeatureswithlargesttotalimportancescoresandsmallesttotalsimilarityscores.Mathematically,thiscanberepresentedasfollows:
In(1),therearetwoobjectives:tomaximizethesumoftheimportancescoresofindividualfeatures,andtominimizethesumofsimilarityscoresbetweenanytwofeatures.wetakeacommonapproachinoptimizationandconvertmulti-objectiveprogrammingtosingleobjectiveprogrammingusinglinearcombination.
2.5Solutiontooptimizationproblem
Theoptimizationin(2)isatypical0-1integerprogrammingproblem.Onepossibleapproachwouldbetoperformexhaustivesearch.However,thetimecomplexityofit,??(????),istoohightomakeitapplicableinrealapplications.Weneedtolookformorepracticalsolutions.
Inthiswork,weproposeagreedysearchalgorithmfortacklingtheissueAlgorithmGAS(GreedysearchAlgorithmoffeatureSelection)1.ConstructanundirectedgraphG0,inwhicheachnoderepresentsafeature,theweightofnode??is??andtheweightofanedgebetweennode??andnode??is 2.ConstructasetStocontaintheselectedfeatures.InitiallyS0=?.
3.Fori=1…t,(1)Selectthenodewiththelargestweight,withoutlossofgenerality,supposethattheselectednodeis
(2)Apunishmentisconductedonalltheothernodesaccordingtotheirsimilaritieswith??.Thatis,theweightsofalltheothernodesareupdatedasfollows.(3)Add??tothesetSandremoveitfromgraphGtogetherwithalltheedgesconnectedtoit:4.OutputSt.
Theorem1:WiththegreedysearchalgorithminFig.1onecanfindtheoptimalsolutiontoproblem(2),providedthat??????,where??denotestheselectedfeaturesetwith??Proof:Thecondition????indicatesthatwhenselectingthe(t+1)-thfeature,wedonotchangethealready-selectedtfeatures.Denote????={????|??i=1,…,t},where??istheki-thfeatureselectedinthei-thiteration.Thenthetaskturnsouttobethatoffindingthe(t+1)-thfeaturesothatthefollowingobjectivecanbemet.
Since,wecanrewrite(3)as
Andsince????and={??|i=1,…,t??},(4)equals 3.EXPERIMENTSETTINGS
3.1DatasetsInourexperiments,weusedtwobenchmarkdatasets.1)Thefirstdatasetisthe.govdatawhichwasusedinthetopicdistillationtaskofWebtrackofTREC2004[28].2)TheseconddatasetistheOHSUMEDdata[9],whichwasusedinmanyexperimentsininformationretrieval[6][10],includingtheTREC-9filteringtrack[26].3.2Evaluationmeasures
3.2.1Meanaverageprecision(MAP)
MAPisameasureonprecisionofrankingresultsItisassumedthattherearetwotypesofdocuments:positiveandnegative(relevantandirrelevant).Precisionatnmeasurestheaccuracyoftopnresultsforaquery.Averageprecisionofaqueryiscalculatedbasedonprecisionatn:
3.2.2Normalizeddiscountcumulativegain(NDCG)
NDCGisdesignedformeasuringrankingaccuracieswhentherearemultiplelevelsofrelevancejudgment.Givenaquery,NDCGatpositionninisdefined3.3Rankingmodel3.3.1RankingSVM
RankingSVMmakesanextensionofSVMtoranking;incontrasttotraditionalSVMwhichworksoninstances,RankingSVMutilizesinstancepairsandtheirpreferencelabelsintraining.TheoptimizationformulationofRankingSVMisasfollows:
3.3.2RankNet
RankNetalsousesinstancepairsintraining.itemploysaneuralnetworkastherankingfunctionandrelativeentropyaslossfunction.Letbetheestimatedposteriorprobability??andbethe“true”posteriorprobability,andlet.ThelossforaninstancepairinRankNetisdefinedas
3.4Algorithmsforcomparison
Ourproposedalgorithmhastwovariants.Welisttheminthefollowingtable.
Forcomparison,weselectedIGandCHIasthebaselines.IGmeasuresthereductioninuncertainty(entropy)inclassificationpredictionwhenknowingthefeature.CHImeasuresthedegreeofindependencebetweenthefeatureandthecategories.wealsoused“WithAllFeatures(WAF)”asanotherbaseline,inordertoshowthebenefitofconductingfeatureselection.4.EXPERIMENTALRESULTS
4.1The.govdataFig.2showstheperformancesofthefeatureselectionmethodsonthe.govdatasetwhentheyworkaspreprocessorsofRankingSVM.Fig.3showstheperformanceswhenusingRankNetastherankingmodel.Inthefigures,thex-axisrepresentsthenumberofselectedfeatures.ExperimentalresultsindicatethatinmostcasesGAS-LcanoutperformGAS-E,althoughnotsignificantlyExperimentalresultsalsoindicatethatwithGAS-LandGAS-EasfeatureselectionmethodstherankingperformancesofRankingSVMaremorestablethanthosewithIGandCHIasfeatureselectionmethods.4.2OHSUMEDdataFig.4showstheresultsofdifferentfeatureselectionmethodsontheOHSUMEDdatasetwhentheyworkaspreprocessorsofRankingSVMFig.5showstheresultsofdifferentfeatureselectionmethodsontheOHSUMEDdatasetwhentheyworkaspreprocessorsofRankNet4.3Discussions
Fromtheresultsofthetwodatasets,wemadethefollowingobservations:1)Featureselectioncanimprovetherankingperformancemoresignificantlyforthe.govdatasetthanfortheOHSUMEDdataset.
2)OurproposedalgorithmsoutperformIGandCHImoresignificantlyforthe.govdatasetthanfortheOHSUMEDdataset.
Tofigureoutthereasons,weconductedthefollowingadditionalexperiments.WefirstplottedtheimportanceofeachfeatureinthetwodatasetsinFig.6.Thex-axisrepresentsfeaturesandthey-axisrepresentstheirMAPvalueswhentheyareregardedasrankingmodels.Furthermore,weplottedthesimilaritybetweenanytwofeatures(intermsofKendall’s??)inthetwo\datasetsinFig.7.Here,bothx-axisandy-axisrepresentfeatures,andthelevelofdarknessrepresentsthestrengthofsimilarity(thedarker,themoresimilar).Basedonthediscussionsabove,weconcludethatiftheeffectsoffeaturesvarylargelyandthereareredundantfeatures,ourmethodcanworkverywell.Whenapplyingourmethodinpractice,therefore,onecanfirsttestthetwoaspects.5.CONCLUSIONSANDFUTUREWORK
Inthispaper,wehaveproposedanoptimizationmethodforfeatureselectioninranking.hecontributionsofthispaperincludethefollowingpoints.1)Wehavediscussedthedifferencesbetweenclassificationandranking,andmadeclearthelimitationsoftheexistingfeatureselectionmethodswhenappliedtoranking.2)Wehaveproposedanovelmethodtoselectfeaturesforranking,inwhichtheproblemisformalizedasanoptimizationissue.3)Wehaveevaluatedtheproposedmethod
溫馨提示
- 1. 本站所有資源如無特殊說明,都需要本地電腦安裝OFFICE2007和PDF閱讀器。圖紙軟件為CAD,CAXA,PROE,UG,SolidWorks等.壓縮文件請下載最新的WinRAR軟件解壓。
- 2. 本站的文檔不包含任何第三方提供的附件圖紙等,如果需要附件,請聯(lián)系上傳者。文件的所有權(quán)益歸上傳用戶所有。
- 3. 本站RAR壓縮包中若帶圖紙,網(wǎng)頁內(nèi)容里面會(huì)有圖紙預(yù)覽,若沒有圖紙預(yù)覽就沒有圖紙。
- 4. 未經(jīng)權(quán)益所有人同意不得將文件中的內(nèi)容挪作商業(yè)或盈利用途。
- 5. 人人文庫網(wǎng)僅提供信息存儲空間,僅對用戶上傳內(nèi)容的表現(xiàn)方式做保護(hù)處理,對用戶上傳分享的文檔內(nèi)容本身不做任何修改或編輯,并不能對任何下載內(nèi)容負(fù)責(zé)。
- 6. 下載文件中如有侵權(quán)或不適當(dāng)內(nèi)容,請與我們聯(lián)系,我們立即糾正。
- 7. 本站不保證下載資源的準(zhǔn)確性、安全性和完整性, 同時(shí)也不承擔(dān)用戶因使用這些下載資源對自己和他人造成任何形式的傷害或損失。
最新文檔
- 2025鋁合金門窗銷售合同
- 2025瑪雅物業(yè)管理合同
- 谷物種植氣候智能農(nóng)業(yè)技術(shù)考核試卷
- 審議公司相關(guān)管理制度
- 審計(jì)行政人員管理制度
- 員工個(gè)人印章管理制度
- 垃圾轉(zhuǎn)運(yùn)現(xiàn)場管理制度
- 電視機(jī)防水防塵技術(shù)與應(yīng)用考核試卷
- 大學(xué)浴室安全管理制度
- 埃及取消閉環(huán)管理制度
- 規(guī)范網(wǎng)絡(luò)設(shè)備管理制度
- 2025年鐵路列車員(中級)職業(yè)技能鑒定參考試題庫-下(判斷題)
- 電商運(yùn)營崗位技能測試卷
- 2025工程建設(shè)項(xiàng)目多測合一成果報(bào)告書范本
- 麻醉科麻精藥品PDCA管理
- 語言習(xí)得神經(jīng)機(jī)制探究-深度研究
- 兒童發(fā)展問題的咨詢與輔導(dǎo)-案例1-5-國開-參考資料
- 2025年河北石家莊市市屬國有企業(yè)招聘筆試參考題庫含答案解析
- 2025年國航股份地面服務(wù)部校園招聘筆試參考題庫含答案解析
- 宣傳物料制作合同范本
- 2025年度安徽白帝集團(tuán)限公司社會(huì)招聘高頻重點(diǎn)提升(共500題)附帶答案詳解
評論
0/150
提交評論