http://www.informatik.uni-trier.de/~ley/db/conf/sigir/sigir2000.html SIGIR 2000 http://doi.acm.org/10.1145/345508.345564 17 Document Centered Approach t o T e x t Normalization aberdeen acquisition adaptation adaptive alembic analysis annotated annual applications architecture association automatic baldwin based beatrice boguraev boundary building burger cache chen church clarkson columbia computational conference corpus darpa decaying description development disambiguation discourse doran eagle editors engineering english exponentially extensible gale general germany guessing hearst hirschman icslp identifying ieee indexing induction information ings intelligence international internationals interpolation june kanfmann kaufman kenneth kuhn language large lexical linguistic linguistics machine macmillan mani marcinkiewicz marcus mary maryland message mikheev mitchell mitre mixtures model modelling models montreal morgan mori multilingual munich names natural newswire nonlinear number pages palmer pattern penn press proceedings processing proper pustejovsky recognition references research retrieval reynar riao riley robinson rosenfeld rule santorini sense sentence seymore sigir signal sixth some speech srinivas system term text topic transactions tree treebank understanding university unknown used using vilain volume wasson word workshop yarowsky http://doi.acm.org/10.1145/345508.345665 65 Bayes about accounts actual again almost analyses annual arbitrary argues assumptions average bayes both bring certainly combination combinations combining combmnz commonly concerning conference constituent convincingly correlated cottrell cross currently data demonstrated dependence derived development documents does eighth elimination emphasize equation essentially evaluation evidence fact finally from full fusion generally given have herman highly hold however improved improvement improvements independence information irrelevant just knowledge knows less likely linear made makes model more multiple needs note number only optimal outputs overview pages partial performance performing perhaps power powerful practice probabilistic proceedings produce proposed pursuing query rankings references relevant requires research retrieval returned robust scores searches second sets shaw sigir similarity some something sophisticated statistics stituent strategy such systems ternational text that these this transition trec unsurprising unsurprlsing untrue used using validation very vogt voorhees which while will within yield http://doi.acm.org/10.1145/345508.345603 37 Does "Authority" Mean Quality? Predicting Expert Quality Ratings of Web Documents algorithms amento analyzing april associated atlanta authoritative automatic baldonado bharat brin bringing browsing buckley card chakrabarti citation collection collections communicates compilation computer conference constructing contextual cornell death department departrnent development diehl digital discrete distillation electronic empirical environment evaluation evolution exploration extracting forager forthcoming francisco from frontier gather gibson hearst henzinger hill human hyperlink hyperlinked implementation improved information interaction interests interface interfaces isdn january kleinberg large lawfulness libraries life management march motwani networks order organizing page pagerank paper pirolli pitkow pittsburgh press proceeding proceedings raghavan rajagopalan ranking references related research resource resources retrieval robertson scatter schank science siam sigir silk sites smart sources stanford structure structures supporting symposium system systems tech terveen text thesis topic topically transactions university usable user vancouver very virginia visualizing webbook wide winograd working workspace world york http://doi.acm.org/10.1145/345508.345562 15 Structured Translation for Cross-Language Information Retrieval access ambiguity american bates bian catalogs chen computational conference contests cross database design different disambiguation editor eleetronze fellbaum information international introduction issue journal june language lexical linguistics model needs online pages polysemy press references research resolving retrieval sczence second seeking sense society special state subject target translation veronis word wordnet http://doi.acm.org/10.1145/345508.345656 59 On the Design and Evaluation of a Multi-dimensional Approach to Information Retrieval advanced algorithms application approach arpa available buckley comm conf conference dasdk data database dimensional document evaluation extended february foundation frei frieder grossman gyssens harman heuristics holmes http indexing information integrating itcc jasis john kimball kluwer lakshmanan language large length library lynch macleod march mccabe microsoft mitra model msdn multi multidimensional nist normalization olap pages pivoted proc proceeding proceedings query references relational rence retrieval roberts salton schauble sdlcdoc september sequel sigir singhal sons space stonebraker structured text textual toolkit trec userdefined vector very vldb warehouse wiley wilkinson with wong yang http://doi.acm.org/10.1145/345508.345636 51 Multimedia Information Retrieval from Recorded Presentations acknowledgements acmmulmnedm additionally applications arons audiovideo authonng auto automated automatic behavior bemd building chnstel communications computer digital discussion documents dunng evaluation existing from generation german gong grudm gupta haskell hauptmann human ieee illustrated implemented information integrated intemctwely into journal learned lecun lessons library like lnteractton logging medla miiller mltiattve mnlnmedla model module muller mulmned mult multimedm multimedta ofacm ottmann part presentation presentations presentatmn proceedings proceedt processing project rabmer recorded recording references relevance replay research retrieval sanocki shahraray skimming society special speech speechskimmer sprmger summarization summary supported synchromzation system systems techniques tele terebyte thank this transactions tssue user vtdeo wactlar which zupanclc http://doi.acm.org/10.1145/345508.345597 34 Topical Locality in the Web aaai accessibility adaptive adding agents aggregation algnnthms algorithm algorithms also amitay analysis analyzing anatomy applying approach architecture ardo artificial associated atlanta australia authoritative automatic autonomous available balabanovic based being belew berg best bharat boyan brin brisbane broder brummer building canada categorization chakrabarti clustering collaborative collection common communications communities compilation computer computing conference content context cornell crawling current data davison dean death department dept description design desire development different discovery discoweb discrete distillation documents dora dreilinger dynamic edinburgh editors eenth efficient eighth eindhoven electronic engine engines enhanced environment environments etzioni evil expanded experiments expert factors features fetuccino finding first focused francisco freitag from frontier garcia geneva genvl gerasoulis gibson giles google grouper guide hcrc henzinger herscovici home howe html http human hyperlink hyperlinked hyperlinks hypermedia hypertext hypertextual identify ieee implementation importance improved indexing indyk inferring information inquirus institute intelligence intelligent interface internalizing international internet jacovi joachims joint jones kaufmann kleinber kleinberg kleisouris koch large lawfulness lawrence learning learns lieberman life link links ljubljana local locality lundberg maarek machine magazine maintenance master mathematics mcbryan measuring melbourne menczer meta metacrawler metasearch methods mitchell mladenic molina more morgan nature neci observations optimizing ordering originally overlap package page pages paper pelleg personal phrasal pirolli pitkow porter portland position post poster prefetching prepared proceedings program project public published query radar raghavan rajagopalan readings recommendation references related relative report reportsld research resource results retrieval review robot rutgers savvysearch scale scaling science scotland search searchenginewatch seattle selberg sept sereport services seventh shaul shoham shtalhaim siam sigchi sigir sigmod size slovenia soda soroka sources specific stefan stripping structure suffix sullivan support switzerland symposium systems taming target technical technique technology telematics text than that thesis through tools topic topical topology toronto tour univ university usermodeling using version wang watch webmasters webwatcher which wide with wiuet work workshop world wwww zamir http://doi.acm.org/10.1145/345508.345632 50 Cognitive Approach for Building User Model in an Information Retrieval Context aaai abased abdallah accept access agent agents aimed allen amalthaea amsterdam analyse analysed analysis annual applciation applying approach april architecture arist artificial available based bayer behavior behaviour behaviours belkin bernard best between billsus borgman catalogs centred change christine clause cognitive communication component components composed conf conference constructing construed context daniel design developing dialog differ discovery documentation documents dublin during ecosystem edwards effort elucidate engage enhance evaluation evolving exhibit experience filtering four france from functional goals graham green hard have identify identifying ifip implications important inferring inforarntion information intelligence intelligent interact interaction interactions interesting interface international internet ireleand jasis journal june know knowledge learning library line lngwersen london lyon machine manage march match methods model models mofita most moukas multi multiagent murarnutsu must national needs ongoing other particular parties pass payne people personnalis perspective plans point potentially pour pracitical procedures proceeding proceedings reason references relevant research resource retireval retrieval review rinformation robust science scientifiques search searching seeking september shinoda sigir simple sites situation some spring stanford strategies structuration structure stuttgart sugar suggest sutcliffe symposium syskill system taylor techniques technology tell text that these they this time together universit user using variety view viewed webert which while whole with would http://doi.acm.org/10.1145/345508.345591 31 Partial Collection Replication versus Caching for Information Retrieval Systems accepted acharyn acknowledgments advanced agency agreement ahamad algorithm algorithms alonso alranar also american amherst analysis annual anotonio application applications april architecture architectures arpa august austin austria authors available award baentsch balance based been benzel berkeley bestavros better both brendon brown bruce building burkowski caching cahoon california callan candela case characterization chicago cluster collection college commerce communication competitive computer conclusions conference congress consensus content contract contributions cook cooperative cost couvreur craswell croft cutting data database databases december defense demand department development diego digital dissemination distributed document doug dublin dynamic ease effective efficient engine engineering ensineering equipment evaluating evaluation excite experiences expert expressed factors fifth findings forth foster foundation france frlfich from full gaithersburg gareia government grant grants greece harding harman have hawking hellas holmedahl hpdc htlp html htndlpaperslp http huang ieee incremental information infrastructure inquery institute intensive interact international internet intia into introducing ireland issued jajodia january journal july june kathryn large legislative lesse level libraries library llwwwsconf lnternatimud load macleod managemem management markatos martin massachusetts material mccoy mckinley mealey miller molin molter naming national necessarily news nineteenth nordin nothing number october office opinions order orleans overview pages parallel parallelizing paris part partial patent performance performing phillips possible practice princeton principles proceech proceedings process processing projects prototyping providing queries quorum ranking recommendations reduce references reflect relevance repfication replica replicated replication report research results retrieval review russell scalable scaling scheme schemes science search searching second selection september server seventh shared shivaratri sigir simpson singhai smaith society software spain spdp special sponsors stan stanford states statistical strategies study sturm support supported survey switzerland symposium system systems tcense technical technology tenth texas text thank that theory thesis this thistlewaite thomas those today tomasic track trademark traffic transaction trec trinity under united university updates uses using utilizing valencia variety vegard very vienna wang wide wilder wisdom with without wolfson wong work workloads workstation world would wwmexcite yield zdonik zeitler zhou zurich http://doi.acm.org/10.1145/345508.345556 13 The Feature aaai academic acknowledgement adding advanced aihara algorithm algorithms allan also although amati amemcan amount analysis another applied approach around atsushi authors based bayes been believe both buckley caraballo carried cases categorization categorzzation center charniak classification classifier classifying comparative comparison complicated components conald concerns considerable consistency consistently conventional corpus course data determining difference different dimension discussions distributed document documents does domain easily editor effect emnlp environment event examination example existing experimental experiments explained exploratory fact feature feedback fields figure fine first freq frequency from funded further good greiff grobelnik have helpful hierarchically hierarchy highly however icml include informatics information institute introduced investigation issue japan japanese joachims journal kageura kenro kita kluwer koller language laplace large lastly learning leaving lewis like llsf logics machine mainly maldenid many martin mccallum measure measures methods mixture model modeling models more nacsis naive national needs nevertheless nigam notes nouns ntcir observed obtained only other outperform overturned paper partly pedersen performance performs phenomena poster press probabilistic processing proe project promotion proposed prtfidf quantitative quantity reasonably recognition reduced references referred relevance remarks representation research resources retrieval reuters rijsbergen rocchio rocehio sahami salton same scale science seems selection semantze sharing showed shown sigir since singer smaller society soczety sometimes specially specific specificity speech standard statistical statzstzcal study such superiority support systems takasu tasks techniques term terms test text textual tfidf tfidfcos than thank that their theoretic theory think this today tokyo tuning tutorial ubiquitous uncertainty university used using utilization values variance variety vector very weighting well wessel which while with wong words work working workshop worth would yang http://doi.acm.org/10.1145/345508.345554 12 L i n k - B a s e d a n d C o n t e n t - B a s e d E v i d e n t i a l I n f o r m a t i o n in a B e l i e f Network Model addison algorithms also american analyzing anatomy annual appear applications associated australia authoritative authority automatic available average baeza based bateman bayesian belief bell better bharat bibliographic bichteler both brazilian brin brisbane california canada cancum chakrabarti challenges clustering cobweb cocitation collection combination combined combines compared compilation compressing computing conference content coupling craswell crawler crestani croft data described designed development discrete distillation distinct distributed document documents dora each eaton edition editors elsevier engine england environment essex evaluation evidence evidential experiments extra ferenee figures forum frakes francisco from gain gains gaithersburg gerard gibson gigabytes golgher grifliths hall harman have hawking henzinger hill hyperlink hyperlinked hypertextual images important improved indexing inference infor information institute intelligent interdocument international introduction isolation jagopalan jansen journal judea kanfmann kaufmann keinberg kleinberg koenig laender large life link luckhurst management managing marion maryland mcgill mcgraw mechanisms method mexico model models modern moffat morgan muntz national neto network networks ninth organization overview page pages pasi pdbeiro pearl performance pieces plat plausible powerful precision prentice probabilistic problem proc processing produced proe provides queries query raghavan ranking rankings real reasoning references regard requiring research resource results retrieval ribeiro rive saddle salton saracevic scale science search seventh shown siam sigir silva similarity single small society soft sources space specially spink spire springer standards string structure structures study switzerland symposium systems table techniques technology text than that their them theory this thistlewaste time topic toronto track transactions trec turtle upper user users using values vector vectorrecall veloso verlag very wesley when which wide willett with without witten wong world wwws wwwt yates yields york ziviani zurich http://doi.acm.org/10.1145/345508.345593 32 Hierarchical Classification of Web Content aaai able accuracy acknowledgments advances advantages agrawal alessio algorithm algorithms analysis annual anonymous application applications approach apte artificial assigns automated automatic automatically based baseline between boolean bringing browsing burges categories categorization categorizing category chakrabarti chen choice cikm classes classification classifiers classifying cluster clustering code cohen collections combining comments companion comparative compared comparison comparisons computer computing conference construe content context could cover damerau data database databases decision decisions decreasing definitely delivery demonstration dependencies development difference directions discriminate document documents dramatically dumais ecml effective efficiency efficient egan elements ellis emnlp empirical english enhancing estimation etzioni european evaluation examination expert exploiting factors fast feasibility feature features fiat fields fifteenth find formative found four fourteenth from fuhr function further future gather generation good grateful grobelnik hartmanna have hayes hearst heckerman help hierarchical hierarchically hierarchies hierarchy horwood http hull human hypertext hypothesis icml improving indexing indications inductive information innovative intelligence interesting international into issues joachims john journal karadi kernel kershenbaum ketchum knowledge koller lage landauer language large larkey learning lesk level levels lewis lochbaum looksmart lustig machine machines management many mccallum methods minimal mitchell mladenic model models more moscow much multi multiplicative murray natural nature nauka needed network networks neural news nonhierarchical notes number obtain only optimization order organizing patents pedersen perspective platt predictive preliminary press problem proceedings processing project psychological raghavan reduce reducing reexamining references relevant remde representations requiring research results retrieval reuters reviewers riao ringuette rosenfeld routing ruiz rule rules russian sahami same scalable scatter schiaffino schiitze schtilkopf schwantner score scores scoring sdair search searching second selection sensitive sequential seventh should shown shrinkage sigchi sigir signature simple since singer small smola some springer srinivasan stage statistical stories study subject superbook support symposium system systems taxonomies text that their theory there third this thomas through thus topic training transactions translation trying tzeras usability using vapnik vector verlag very vldb weigend weight weinstein weiss wiener wiley will with within without words work working workshop worth yang http://doi.acm.org/10.1145/345508.345546 6 A u t o m a t i c G e n e r a t i o n of O v e r v i e w T i m e l i n e s alan allan allen alternative amherst analysis annual applications applied approach association athens august automatically back based bikel broadcast broglio browsing california canada carboneu card center chapman chicago cikm city computing conference conrad constructed contingency croft crow dagan darpa databases david demonstration design desktop detection development digital discovering discovery document documents dublin editing editor editors eighth english eric event everitt expansive extracting extraction factors feature features february feldman fertigand fifth finder first francisco freeman from furuta future galaxy gelernter greece hall herndon high histories human implementation information ingalls intelligent interaction interactive interface interfaces international itself james japan jinxi jock john journal july kaehler kansas kaufmann keyword knowledge kumar landscapes language languages lantrip large lavrenko learning libraries life lifestreams line lines london machine machinery mackinlay maloney management marchioni massachusetts metadata metaphor milash miller missouri morgan name natural nevada news november nymble object oopsla oriented pages papka part pennock pennsylvania performance personal pierce pittsburgh plaisant pottier practical proceedings processing programming readings references relationships rennison report research retrieval retrospective review richard robert ronen rose russell schur schwartz scott sdair selforganizing semantic sets shneiderman sigir significant sigplan smalltalk soergel software spatial speech squeak story stuart studies study support swan symposium system systems tables tagger technical technology text textual these think third thomas thompson time timeline timelines timemine tracking tsukuba uist understanding university user using vancouver varying vegas vijay virginia vision visual visualization visualizing visulalzing wallace weischedel widoff wise with workshop written yang http://doi.acm.org/10.1145/345508.345660 61 Learning Probabilistic M o d e l s of t h e W e b about addition advantage aggregate algorithm algorithms along also american analysis annotation annual approach approaches authoritative automated based berger berkeley bharat california ceedings characterization combine combines comparison conclusions conference contain content contrast created croft currently curved decomposition deerwester dempster derive developing development directly disambiguate discrete distillation document dumais empirical engine environment environments evaluation extracted factors figure fitting from furnas future gines hall harshman have henzinger hits hofmann hyperlinked icml identified improved incomplete indexing information international jaguar journal keywords kleinberg lafferty laird landauer language latent learning likelihood link links lists lost machine manner markov maximum method methods mining mixed model modeling models most multinomial natural nature noise occurrences offers order other pages pereira ponte post predictions predictive presented principled probabilistic probability proc procedure proceedings processing quantitative query references related research result retrieval returned royal rubin saul science search semantic semi siam sigir significance society sources standard statist statistical subfamilies such suggesting symposium systematic term terms that tool topic topics translation ulysses urls used useful using validate with words work working http://doi.acm.org/10.1145/345508.345602 36 Incorporating Quality Metrics in Centralized/Distributed Information Retrieval on the World Wide Web' aaai about acmsigir acmtransactions adaptive agents altavista american amsterdam analysis anatomy annual appear applications approach april architecture argus articles ation august automatic award based bases beyond breakout brin browsing cafi callan cambridge career categories change ciolek clearinghouse cnet collection collections computing concep conference cooperative corpus create criteria croft crowder cyberspace data database degraded direct directhit distributed disw dynamic engine environment epscor expansion experi expert exploration extension first from fusion ganch garcia gathering gauch gerhard google gravano groupl gupta harding help html http hyperlink hypertext hypertextual index indexing indexingworkshop inform information inquery intelligent international internet ittc jasis johnson journal july knowledge kral kretser lagoze laird lake largescale learning long lycos machine magellan maghelp management mapping march massachusetts mckinley ments merging metadata methodologies mofffat molina multilingual multiple network networked nicholas number obiwan ontology paepcke page pearce philadelphia point press pretschner privated proceedings project properties proposal protocal public qltydefinitions query rachakonda rafsky ranking ratings references relevance report reportoutcomes resource resources retrieval routing science scout search searching selberg selection september shimmin shoot sigir site society sources spain standards stanford statistical strategies symposium system systems tahoe telltale text towell tual twelth ukans university using valencia vector voorhees voting wang wide wisc workshop world wwwdb wwwvlpages zdnet zobel http://doi.acm.org/10.1145/345508.345541 3 A NOVEL METHOD FOR THE EVALUATION OF BOOLEAN QUERY EFFECTIVENESS ACROSS A WIDE OPERATIONAL RANGE academic acta algebra algorithmic algorithms american analysis annual another approaches april arlington arnold arofonsky artificial asis aslib august automatic belkin between blair boolean butterworths case characteristics cleverdon cliffs cognitive combinations comm communications comparison computer concepts conference consistency conventional cranfield croft cross current data databases deductive devices doctoral document eaglewood effectiveness electronica elsevier evaluation exact expansion experiment experimentation experiments fidel first frants freeman freestyle full gaithersburg game granum guildford hall harman harter hatter hersh heuristic hickam hill http iivonen implementations index indexing information ingwersen institute intelligence interactive international introduction isbn issues john jones journal july kantor keys knapsack kristensen laaksonen lancaster language libri linear logic london look management march maron martello match mcgili mcgraw mckinin measuring medical medlars medline method methodology model modern national natural newell nist online operations orlando overlap paris partial performance perspectives pragmatics prentice press pricai problems proceedings processing programming progress project proposed publ queries query range rapid references research resources retrieval retrieving review revisited saiton salton saracevic science search searcher searching seeking selection seventeenth shapiro sievert sigir singapore smart society sons sormunen sparck spec springer standards state structured study sutcliffe switzerland system systems tague tampere tamperensis techniques technology term terms testing tests text textbook textretrieval thesis tibbo today tool toth trec turtle universitatis university velin verlag warner washington wide wiley willett williams with workshop york ziidch http://doi.acm.org/10.1145/345508.345543 4 Evaluating Evaluation Measure Stability acsys alistair allan analysis annual april august australia automatic autonomous available bailey blustein bruce buckley butterworths callan chapter charles chart chris christopher clarke classification cleverdon cliffs collections comparisons conference construction cooper cormack cornell cranfield craswell croft daniella data david determining development document documentation donna edition editor editors edward effectiveness efficient eighth electronic ellen england englewood evaluating evaluation experiment experimental experimentation experiments factors fang feng fidel filtering fourth gerard gordon grunfeld haircut harman hawking html http hull implementation indexing information ingwersen inquery international issue james jamie janet january jean jersey jones journal judgments justin karen kaufmann keen kwok large lesk lewis malin management mayfiled mcnamee measure measurement melbourne michael mills moffat morgan nick nist october optimizing overview pages palmer part performance peter piatko pircs pragmatics prenticehall presenting press proceedings processing publication pubs query raya readings references relevance reliable research results retrieval revisited rijsbergen ross salton scale selecting seventh sigir sixth smart sparck special state statistical sutcliffe system systems tague ternational test testing text third track trec using vanrijsbergen variations version volumes voorhees walz wilkinson willett william williamson york zobel http://doi.acm.org/10.1145/345508.345598 35 Interactive Internet search: Keyword, directory and query reformulation mechanisms compared adcs analysis anick annual assistant assisted australian baddeley bars bateman bruza capacity centre cognitive collins computer computing conference confidence data dennis department design document documentaires easy empirical engine engineering ergonomics error etudes expansion feedback figure formulation forum function generating google harper hautes henzinger human hyperindex iihib imposed improve information informatique international internationales internet intervals iterative jansen journal large lexical life lnternet load made marais mcarthur measure mechanisms michon mihalcea moldovan moricz operators pages paraphrase perceptual performance proceedings psychology quarterly queries query randomization real references refinement regularity relation relevance represent retrieval riao saracevic science search searches searching seeking sigir silverstein spink spinks study subjects sydney symposium tapping term terminological that third tipirneni university user using very visited wickens wide wordnet world yahoo http://doi.acm.org/10.1145/345508.345582 28 An Investigation of Linguistic Features and Clustering Algorithms for Topical Document Clustering aberdeen alembic algorithms analysis annual applications august baeza based bates broadcast burger cliffs cluster cohen common compounds conference croft darpa data databases description detection development doddington domains editors englewood evaluation february finding fiscus frakes garofolo gather groups hall hearst hemdon heterogeneous hirschman hypothesis information integration international interpreting introduction jersey june kaufman management march marti martin message mitre news nist nominal nonlinearregressionanalysis pages pedersen prentice proceedings processing queries reexamining references research results retrieval robinson rousseeuw scatter sigir similarity sixth slgmod structures system textual topic tracking understanding used using vilain virginia watts wiley without workshop yates york http://doi.acm.org/10.1145/345508.345573 22 D o c u m e n t Filtering M e t h o d Using Non-Relevant Information Profile achieve adaptive analysis approach automatic based choi clarit conference conservative contribution description document evans existing expansion experiments feedback filtering gatford hall hancockbeaulieu harman hashimoto high hindle hoashi hull information inoue jansen jones lewis many matsumoto method nist notes okapi optimization overview pereira performance prentice proceedings processing published query references relevance retmeval retrieval robertson rocchio roma seventh sigir singhal smart stoiea system systems take text third track trec voorhees walker word zhai http://doi.acm.org/10.1145/345508.345648 56 Theme-based Retrieval of Web News access accuracy advances allan along applications approach approaches archibald architecture ariadne articles arucles august average behavior bowman brown building built burges busi carbonell categories categorization classification classifier conference confidence conj culture cyclic danzig decrease degradation detecting detection development digital discovery distributed ecdl effect effects european event events evident examination examples feature ferreira figure final first force from gaspar grilo hardy harvest higher however ieee information informatton intelligent international january joachims july kernel kopf large lavrenko learning library like line linear lssue ltbrartes making manber maria measured methods mimmize mimmizing models negauve news number obtained other pages papka pierce politic politics positive practical press proceedings reduced references regression rence research retrieval retrteval scale schwartz score seasonality second selected selection sharply sigir silva smola special sports strategies strategy such support system systems tcmdency tends text this time topics tracking trained uniformly vector were with yang year http://doi.acm.org/10.1145/345508.345661 62 Effects of Out of Vocabulary Words in Spoken Document Retrieval* advanced allows analyses appear approaches auza basic cambridge caused clearly compensate computer different document errors expansion figure garofolo irck johnson jones jourlin laboratory level lower more much overview pereira performance proc proven rate reach recognition references report results retrieval robertson rrckjones seen sets setups shown sigir simple singhal speech spoken summary system technical techniques test that this track transcription trec university using various voorhees with woodland words http://doi.acm.org/10.1145/345508.345652 58 Exploration of a Heuristic Approach to Threshold Learning in Adaptive Filtering adaptation adaptive additional algorithms among appear aspects available beta better between built calibration case changes chengxiang chunks clarit closerto conclusions conference conservative crucial current currently david delivery demonstrated different disabling disastrous docs draw during editors effect effectiveness eighth emilia empirically enough evaluations evans eventually examined examples experimented experiments extend figure filtering final fixed following formally framework from function further gain gamma general generality government govlpubsltrec grot harman higher highest html http hull incurring information initial initially interesting jansen lead leads learning lltrec long loss lowering main many mean method methods model most necessary nist norbert observable office only operational optimal optimization other otherwise outperform over penalty perform performance perhaps period peter poorly possible printing problem proceedings publication pubs ratio references reinforcement relating report result resulting retrieval robertson roma runs same scoring seen setting settings seventh several shows special steve stoica such system temporary term text that their theoretical this those threshold thresholds time track trade trec unit updating utility variables varying vector voorhees washington ways where which while with within work zhai http://doi.acm.org/10.1145/345508.345643 54 Ranking Digital Images Using Combination of Evidences account addison adjust adopting allows also amati another applications applicatmn approach april ashley assign assigned associated australia based best between built captures carnap certain certainty chang chicago classical collection colloquium color combination combining company components computed computer concept concepts conceptual conference confidence content context correspond currently data database databases deduced degree described dexa different documents driven electronic engineering evaluating evidence exchange expert expressweness extended fact factor field figure finding first flexible flickner florence formalism formula forrnahsm foundations frequency from genericlty given glasgow gorkani graph graphs hafner hence here hierarchy hong huang ieee image imaging importance indeed index indexed indexes induction information inforrnatlon international into irsg italy january jose journal july knowledge label level like linking logical logm loth machine manipulation media melbourne mentioned methods mind mpact multimedia multiple niblack object objects occurrences order organise oriented ounis parameter parameters pasca pentland people petkovic photobook photographs picard piece position press probability proceedings processing properties propose provide publishing qbic query ranking rapidity references refined refining relation relevance relevancevalue relief relies represent represented retrieval rodden rvscg rvsi rvsx sawhney schema scheme schemes sclaroff seattle semantics september sigir singh single smith sowa specificity spie stelle stprage structures subgraph subgraphs symposium system systems taipei takes techniques that their this tools touch tree uncertain university used user using value vary video vision volume weight weighting weighung well wesley where which will yanker http://doi.acm.org/10.1145/345508.345625 47 The Effect of Query Type on Subject Searching Behavior of Image Databases: an Exploratory Study aaai access american analysis annual arbor archives armitage attributes available based bellardo berkeley besser caivl challenges communications conference content context corneli database databases degree design digital dissertation document documentation edie electronic engr enser evaluation experimental fidel hahn high howard http hypermedia ieee image images impact implications indexing information inforrnation inquiry intelligent international investigation jain jorgensen journal july keister korfvidal learned libraries library lmage lnformation lnto management master maybury medford menlo multimedia need organization park pennsylvania performance philadelphia philip pictorial practical press project queries query rasmussen raya references retrieval review scalable science services smith society spontaneous spring subject systems task taxonomy technology text thesis trendy trudi types university video visual washington workshop yamashita yapp york zick http://doi.acm.org/10.1145/345508.345512 0 RELEVANCE AND CONTRIBUTING INFORMATION TYPES OF SEARCHED DOCUMENTS IN TASK PERFORMANCE aaai able ablex about academy accounted actions addition aibrechtsen also analysis annual anomalous articulate asis aspects assessing assessir attorney background barry based basis bateman became because beginning behavior being belkin between calif canadian case cases categories categorization changes citing clear clearly clues cognition cognitive complexity conceptualizations confirm conjbrence connecting connection consequently construct contain context contributing contribution contributive could criteria criterion decisions decisive decreased defined definition degrees depends design determinations developed dies different directional document documentation documents does domain during dynamic early efficient eisenberg elements empirical encyclopedias essence europe examining experience experienced expert expertise exploratory explore extent facets feltovich finding finland finnish flows focused ford found framework frameworks from further gavin general great greisdorf guidance hakala hall hand harter have help highly hoffman horizon human importance include indices inference information initial inside integrating interaction jasis journal judge knowledge kuhlthau least library links literature lngwersen logical london longitudinal machine management meaning medford medical meeting menlo mental methodological model modeling models more most mostly much nature need needed nilan norwood notion occur only other over paris park patel performance perspective perspectives phase possible prentice press problem proceed proceedings process processing project provide psychological psychology publications quarterly ramoni rautio reading reasoning reexamination refer references regions relevance relevant representations request research researcher respect results retrieval review reviews riao rland role saracevic schamber science sciences scienece search searchers searching seeking seekmg seems sees select selection september shape signal signaling significant situational social soergel sought sources speaking spink stages state states step structure structured struggling students studied studies study subject subjects such suggest suhonen suomalainen support sutton system systems tactics task technology tentative terms textbooks that their them theory these they thinking thirds this thus tietovirrat today tools topic topicality toward towards traits tutkija type types understood unstructured useful usefulness user users vakkari various varying wang well were what when white williams with would yhteiskuntatieteiden http://doi.acm.org/10.1145/345508.345572 21 Text Filtering by Boosting Naive Bayes Classifiers aaai accurate active adaboost advances algorithm algorithms also amit analysis annals applied autonomous bagging barlett based bayes bayesboost because before belkin best boosting breiman buckley calculating callan categorization chan classification classifters coin collection combined communications compared comparison competitive conclusion conf conference confidence consideration contrast cool cortes could croft data david decision description despite determining dinstl disadvantage disadvantages document documents does drucker effectiveness employing entries evaluating examination example examples experimental experiments explanation fact features feedback figures filtering finally first following forms freund functions further future genetic given grunfeld harman have however hull icml improve improved information into joachims kwok learning length lewis linear machine make makes many margin mccallum measure mentioned methods might mitra model moderate more naive neural nigam normalization number observed opposed optimization optimize optimizing other outperforms overview papka performance pirc pircs pivoted pool positive predictions predictors prepares probabilistic probabilisticanalysis probability proc processing proe queries query quinlan ranked rated ratios references related relevance result results retrieval retrieved rocchio routing runs salton same schapire second seems show shows sides sigir significantly singer singhal slgir slightly small sparse statistics still stumps submitted suitable systems take text than that their this threshold track training trec trees unjudged unlikely uses using utility various very votes voting weight weights well when with work worse yang yoram zone http://doi.acm.org/10.1145/345508.345638 52 Influence of Speech Recognition Errors on Topic D e t e c t i o n added allan annual auzanne bnasr broadcast carbonell cdet clusters communication conclusion conference contemporaneous darpa degraded detection development dharanipragada document doddington domain errors european evaluation expansion experiments february figure final find fiscus franz garofolo harman have information international lund manually martin mccarley mixture news newswire nist over overview particularly pereira performance performed phase pilot plan proceedings ranked recognized references report research results retrieval roukos segmentation sets sigir significantly singhal small speech spoken stanford story study system technology test that three topic track tracking transcribed transcription trec understanding unlike version voorhees ward with workshop yamron yang http://doi.acm.org/10.1145/345508.345667 67 Finding Relevant Passages using Noun-Noun Compounds: Coherence vs. Proximity about adam adams after algorithm ally altavista analysis annual appropriate argued arises artificial assess asuncion atolingua average averaged based basic before behalf benjamins bruce build calculated california candidate case categories ceil chance chaos coherence coipns collaborative column combining comparison compounds computing conceptual conceptually conclusion conditions conference confining constructed construction content continued coolen corpus correlated could croft david defined depend distribution documents dramatic each editors eduard effective effort elitabeth ellen engine engines english european evidence example experience experiment expert farquhar fikes filter formation fornx found fredrie from further future gain gardiner given gives gomez have hearst hence here hoenkamp http ieee illters important improve increase indexed information input integration intelligence international internet into introduction invariant irrelevant isolated james john know knowledge laboratory lacnnae lambert last level lexical liddy linguistic link lnternet london longman major marti mary match matched matches matching material merits methods mihalcea modeled moderm moldovan myaeng near need nici nijmegen nist nominal nonparametric noun novel number observations occurred oecurs onno ontological ontologies ontology operator operators ordered original over pages paired parametric part pass passages passed perez potentially pratt precision press proactive problem proceedings processing produced promising proximity queries query question received references relations relevant report represented results retrieval retrieved rice richard riedl riet rreported ryder sample scored search searches second sehomaker select semantic shows sigir significant simple slagle solving spectrum spotting stanford stegemar subset summarized sung supporting syntactic system table tailed technical test text that then therefore these thesis third thirteen this those through tong tool total transformation translate translated trec type types under university update user users using valerie volume wanda wdcoxon where whole wilcoxon with word wordnet work workshop http://doi.acm.org/10.1145/345508.345594 33 A Practical Hypertext Categorization Method using Links and Incrementally Available Class Information aaai advantage against algonthms algorithms also analysis analyzed andersen andrew annual areas articles august australia authoritative automated available based bayes bayesian become been bharat both byron caia california carefully categorization category chakrabarth chakrabarti changes chidanand claims classes classification cleaned collection compared comparison comparisons computer conference consisting constrained content contributed contributions conventional coverage craven croft damerau data david dayne designed detailed developed discrete distillation document documents dora effectiveness effectivness efficiency encyclopedia enhanced environment event examination experiments explored extended extract factor factors flexibly framework francisco fred freitag from fscore gained generalizable gibson hayes hearst help henzmger http hyon hyperclass hyperhnks hyperlinked hypertext ieee improved improvement improving included increase independent individually indyk information internal international iral jeong jung kamal kind kleinberg klemberg knowledge known kumar kyun language learning lewis limitation link links lowered machines made make making management mann marc mark mart mccallum mccauum melbourne method methods mining mitchell models mook more myaeng naive neighborhood niernhurg nigam nigram only over overview pages pasquo positive prabhakar previous proc processing project proposed provide ptotr publicly raghavan rajagopalan references repetition report research retreival retrieval revealed ringuette same savoy scheme schmandt school science score scores sean searching second september shell sholom shorter showed siam siattery sigir sigmod significant similar since situation soumen sources sridhan standard strategies structure subject such sufficiently suitable summarization sung support symposium systems table take target tasks terms test text that them theo thoroughly through tomkins topic towards training turtle unavailability used using various vector weis were what while wide wish with work workshop world wwkb yang yiming http://doi.acm.org/10.1145/345508.345666 66 Information Access for Context-Aware Appfiances adaptive agent application applications augmented aware based belkin bovey brown cellular chen coin communications computer conference context croft editor environment filtering from gatford hancockbeaulieu harman hierarchical ieee information jones journal laboratory location mann marketplace memory mobile model network okapi pages personal prediction proceedings programming protocol references remembrance retrieval rhodes robertson same sides strategies system technologies text third trec walker wearable wiley wireless with http://doi.acm.org/10.1145/345508.345668 68 Semantic Explorer TM - Navigation in Documents Collections actions adjustable advantage after also annotated applications athens automatic banners based basic bear both browser built capable capturing categories choice citation classroom clients closely collections commercial communicate compare comprehension content copies copy daily database databases digital distributed document documents each easy explorer export extraction facilitates filtering first form from full gradually granted greece hard have hierarchy information interactive interested issue java keywords large learning lists made make maps more navigational neurok newspaper newspapers notice numerous only organizing otherwise page part particular parties permission personal post preferences prior procedure product products profile profiles profit provided provides proxima pure quick receive redistribute regardless related represented representing republish requires result retrieved search self semantic separate server servers service showed sigir similar software sold space specific structure such system targeted technology term terms that their thematic then there third this those thus tool tree used user users vector vectors well which without work http://doi.acm.org/10.1145/345508.345559 14 I N S Y D E R - A n I n f o r m a t i o n Assistant for Business Intelligen'ce aaai aamas academic access addison advanced agent agentensystcm agents ahlberg aiamitos alertbox amalthaea annual anuary anwendungen applications approach arisem artificial august autonomic autonomous baeza bclkin bedeutet behavior belkin berfin beschaffung bilger billsus bisolutiotff breaking brenner btrf business byrd card case catalog categories clarifying clustering clusters combined computer concept conference conffdata conj coordinating crimmins croft csdl cugini data database debates demonstration design development deystaffweb dienste digital dill direct discovery distribution dlib document dordrecht dubfin ecdl effectiveness eichmann eines elec electronic elektronische elpub endversiongm engine engineering entwicklung environment ethical etzioni european evaluation evolving excerpts excite expansion expert exploration explore extraction farming feasibility feedback filtering find first florence framework francisco frankfurt ftsinformationen full gaithersburg gershon gesch gesichert government graphical grundlagen guest guide hackathorn hall handschuh harman hearst heidelberg henninger hlmi hockley html htnd http ichmann identifying indcx informaffon informatik information informationelle informationsassistenten informationsm informationssuche infovis intelfigent intelhgence intelligence intelligente inter interaction interactions interactive interesting interface interfaces international internet ireland issues italy ivee jennings jlwww jlzhadum journal june kann kaufinann kaufmann keim kluwer knowledge kocnemann konsequenzen konstanz kuhlen language laskowski late lexical libraries library london mackin maes magazine manipulation mann march maryland media model modern moore morgan moukas moux mubler multi multiagent multiple mussier mutamatsu national natural navathe navigating ncsa nielsen nirve nist north novemberidezember nusa october oder offenen office organizing paper papers pazzani pers piatko pisa pollock portland practical practice prentice press printing proc procccdings proceedings process publication publisher publishers pubs pyle query querying rcitemr readings redmond references reiterer relations relevance rence report research results retrieval reuieval review ribeironero rkten robertson roboter rostock saire salre search searches searching second semantic semar september seventeenth seventh shneiderman shueiderman sigir sigit sites smeaton snap society soflwareagenten software space special spektrum springer strategies strzalkowski study suhrkampverlag supplement swinger symposium syskill system systems tamu tanber tech technical technology tenth term text thanos theory think tilebars together topics trec turau uicd uiuc ukldailpeoplelmikewlpubs unipaderborn university untersttitzung useit user using vcerasa veerasamy veerusamy verlag vertrauen vichi views vision visualization visualizing voorhees wagner warehouse washington webbook webert webforager werden wesley what wide will wills wistrand with wittig wooldridge workshop workspace world wrong wwwdb yates york zamir zarnekow zing http://doi.acm.org/10.1145/345508.345615 41 Generation of User Profiles for Information FilteringResearch Agenda Tsvi Kuflik and Peretz Shoval Information Systems Program, Department of IndustrialEngineering & Management aaai agent agents annual applications arttficial assist atlanta august automated autonomous balabanovic based belkin billsus browsing butterworths center chap classifiers clis cognitive coin collaborative combines comm comparision computer computers computing conceptual conf conference content croft decision development digital distributed document engineering environments etzioni eval experiment experimentation experiments expert fact fiction filtering filters forecast framework freitag gatheringfrom georgia guide hanani heterogenious hill hull human ifeeeexpert indusrtial information integrated intelligence intelligent interact interface international introduction joachims joint jones journal kaufman learning leggett letizia libraries lieberman machine madison maes management marchionini mcgill mcgraw methodology mitchell modeling modern moghrabi monlreal morgan network neural norwegian oard ofljca oftheacm overload pazzani pedersert personalized problem proc proceedings processing recommendation reduce references report repreentations research retrieval robertson routing salton same sanchez schutze services shapira shoham shoharn shoval sides sigir sociological sparks spring stereotypes support survey symposium system systems technical text that through tour user users weld wide wisconsin with work world york http://doi.acm.org/10.1145/345508.345624 46 Modeling Question-Response Patterns by Scaling and Visualization 1 accuracy american analysis answered appears applied both cannot concluding conference considerably corporation data difficult dispersion economy effort eighth evaluation experiments figure fitzpatrick form future greater hardness harman implementation information intel item journal judged measurement method model multidimensional must myriad nist noted past performance press process provide provided psychological publication query question questions rather reckase recovery references reflection remarks response retrieval rorvig scaling science sets shape shortcoming similar situations society special studies study supported text that theory these this topic trec used visual voorhees which with http://doi.acm.org/10.1145/345508.345674 73 TimeMine: Visualizing automatically constructed timelines alan allan annual applications athens automatic back cikm city conference development eighth extracting features from future generation greece information ingalls international itself james john kaehler kansas knowledge languages maloney management missouri november object oopsla oriented overview pages practical proceedings programming references research retrieval russell scott sigir significant sigplan smalltalk squeak story swan systems text these timelines timevarying wallace written http://doi.acm.org/10.1145/345508.345646 55 Collaborative Filtering and the Generalized Vector Space Model achieve algorithms allows also american analysis analyzed applications applying artifical artificial august automating based between bradley breese bridge brown buckley carbonell carl chris collaborative common communications comparative comparison computed computer computing conclusion conference content corpus cross danny david denver department different documents duke empirical evaluation exists fact factors feedback filtering form formal fourteenth frederking fully generalized geng gerard gordon grouplens gvsm have heckerman here herlocker however human idea identity ijcai improving information intelligence interests international jaime japan jiang john joint jonathan joseph journal july kadie kaufman konstan language littman madison maes maltz march methods michael miller model more morgan mouth nagoya needs news pages pattie performance plan predictive proceedings profiles ralf ratings references relationship relevance report representations result retrieval riedl robert salton science shardanand shown similar similarities similarity social society space systems technical that traditional translate translingual true uncertainty university upendra usenet user using vector word work yang yibing yiming http://doi.acm.org/10.1145/345508.345628 48 The Role of a Judge in a-User Based Retrieval Experiment across again allan although annual answers aspect aspects aspectual assessor assessors august australia average been bell better between causes changed clarke cluster collections comparing comparisons compressing conclusion conference consistent constrction cormack correct corresponding csiro definitive description development difference digital distribution document documents dominating each easily effectiveness efficient embedded eric evaluation examining expended experiment experiments explicitly expressed figure following france from fuller gather gigabytes have hearst here hypothesis identified images indexing indirect information inlbrmation interactive interface interfaces international james judgement judgements kaszkiel lagergren large list long lowest maid managing many march marti martin matrix measurement measures melbourne michael mingfang missed moffat nearly none nostrand november observed other over overall palmer part paul pedersen phenomena played popular pressure principle proceedings processing provided rate reexamining references reinhold relationships relevance research respectively results retrieval role ross russell saved scatter seventh showed shown sigir sites speech standard still subject subjects swan switzerland systems taken test text that them there they though time topic topics trace tracks trec unclear under using variations visualisations voorhees well were when wilkinson windows witten workshop zurich http://doi.acm.org/10.1145/345508.345576 24 Bridging the Lexical Chasm: Statistical Approaches to Answer-Finding aaai aaal advantages algorithm algorithms allow american analyses analysis annual answer answering answers application approach approaches appropriate asked aspect associates automatic based believe berger berkeley best better between beyond biron both bridging brown buckley building burke candidate cape cases characteristics chasm chicago cocke cognition collection combination combine combinination combining computational computer conclusions conference control cost could croft customer data datasets deerwester della dempster department depending described desired development dialogues different discuss discussed document domain dumais each effectiveness efthimiadis either emerging empirical employed enabling erlbaum expansion experiences experiments faqs fikely files finder finding four frequently from furnas gartner gartnergroup global group hammond harshman here highly hofmann incomplete incorporate indexing individual information jefinek journal kulyukin lafferty laird landauer languages latent lawrence learned lehnert level lexical likefihood linguistic linguistics local lytinen machine management many maximum mercer methods models more multiple november nual okapi ones optimized overlap paper parametrized particular performance performed pietra potential presented presents press probabilistic problem proceedings process processing produced quantitative query question questionanswering questions quite ranks real reasons references related report require research response retrieval roossin royal rubin salton science second seemed semantic simulation situations size society statistical study such system systems tech techniques term text than that there this though tomuro training translation trec ucla unable underlying university usenet using vocabulary weaver weighted weighting weights well were with work world http://doi.acm.org/10.1145/345508.345574 23 Question-Answering by Predictive Annotation aaai anlp annotation answering answers approach april aronson austria basic bell brown burke byrd categories chaudhri choi chong coden cognitive collection communications compressing conference craswell customization damerau database decision disarnbiguation documents editors eighth encyclopaedia ends english evaluation expansion extracting extraction fall falmouth fikes francisco from front gaithersburg gigabytes guru hall hammond harman hawking hill htrnl http identifying images indexing information institute intelligent introduction kaufmann klagenfurt kulyukin kupiec language large lexical lexicalsemantic line linguistic managing mcgill mcgraw message miller modern moffat morgan murax names national natural nldb north nostrand november objects ofanlp official ofsigir online organization overview pages papers prager predictive prentice press problems proc proceedings proper psychology publication published qatrack query question questions radev ranking ravin references reinhold relations report research retrieval robust rosch salton saran seattle seventh sfihari sigir singhal site sixth solutions some special standards support supported suspected symposium syst system systems technical technology text thistlewaite track trans trec turban understanding using very voorhees wacholder washington witten wordnet york http://doi.acm.org/10.1145/345508.345623 45 Lexical Semantic Relatedness and Online New Event Detection allan application broadcasting cambridge carthy chaining chains coherence cohesion cohesive comprehension computational darpa database delaware detection document domain english final flood george graeme halliday harmony hasan hatch hirst indicator international irsg issue james jane journal lexical lexicography linguistics longman march miller morris newark news nicola novel online paper paula pilot proceedings propose reading references relations report representations special stokes structure study summary text thesaural this topic tracking transcript understanding within wordnet workshop http://doi.acm.org/10.1145/345508.345577 25 Building a Question Answering Test Collection aaai access adam albert analysis annual answering appelt approach articficial artificial asked assessments august australia available barbara bases bear boris breck burger burke called case ceedings chicago cohen computer conference darpa david dawn department description development downey draft edie editor editors eighth electronic ellen encyclopedia english eric evaluation experiences explorations fastus ferro files finder five forum frequently from fundamental gunning hammond harman high hobbs holland house htrul http indeueet information intelligence international israel issue january john jones julian june kameyama katz kaufmann kehler knowledge korfage kristian kulyukin kupiec language laura lectures lesk light line linguistic lingusitic lisa lunar lytinen magazine management mani marc martin melbourne message morgan murax murray myers natural nist nonko north notebook november nual pages paper paul pease performance peter plum presented press proceedings processing project pubs qanda question questions rasmussen references relevance report research results retrieval robert robin robust rocks salton schoenberg schrag science scott sentence shapiro sigir sixteenth sixth special spnng starr steven storage structures stuart studies study symposium system systems technical technologies terry test text tice tomuro track trec tyson understanding university usability used using version vladirnir volume voorhees webber wide wiley willett winograd winter with woods world york zampolli zprise http://doi.acm.org/10.1145/345508.345618 42 Variance Based Classifier Comparison in Text Categorization abstract abstracts academic account accui accuracy acwhere affects algorithms also amount analysis annual anwe applied articles average based basis belongs between binary both bound categonzatmn categorization chebyshev chines chio chose class classes classification classifier classifiers classifying comparison conference conferences considered consists construct contain containing corpora criterion curacy data database decompose defined depending derive derived determining development deviation discussed distribution document docurespectively each equation equations especially european examinationof example experiment experimental feature features figure figures first fisher following former frequency from function future gain gausian have hierarchically however inequality information international into inverse japanese joachims kernel koller large larger latter learning least lewis lower machine made make makind many matrix ments method methods nacsis negligible normal nual number obtain order other outperforms paper performance plan plots point preliminary prepared presented problem problems proc quantxtative racy radial randomly ranging refer references regions relationship relevant reliability remmnmg represented research respectively result results retrieval ringuette rocchio rocdata rocrn sahami same select selected sets show sigir similarly size small smaller societies society sored space spon sponsoring stand standard stands study subset subtracted sufficiently support symposium take tend term test text than then these this those tlaht training tramcovariance trieval used using variance varies vector vectors very view weighted were when whether which while will with words yang http://doi.acm.org/10.1145/345508.345658 60 SWAMI: A Framework for Collaborative Filtering Algorithm Development and Evaluation algorithm algorithmic algorithms architecture aste august automating based bergstrom berkeley boosting botchers bottom clustering collaborative combining computer conclusion conducting conference contributions correlation creation demon designing development eigen eigenvector ettlcient evaluating evaluations figure filtering framework freund future goldberg grouplens herlocker human iacovou information informatwn interaction invited iyer jester john konstan largest learning machine maes main method mouth netnews november open overview pearson performing portable preferences proceedings research resnick retrieval riedl scalable schapire shardanand singer smallest social sons standardized statistzcal supported sushak system systems talk theory vapnik wiley word work http://doi.acm.org/10.1145/345508.345579 27 Latent Semantic Space: Iterative Scaling Improves Precision of Inter-document Similarity Measurement Rie Kubota Ando analysis ando approaches automatic banks bartell based bayesian behavior belew blind boguraev byrd case ciples comneff computers conference content cottrell data database deerwester digital ding discourse document documents dumais effectiveness ences encyclopedia explanation external factor free from furnas harshnaan hawaii hicss hill ieee improving indexing informatie information instruments international introduction irregular journal kruskal laham landauer large latent learning manof mantic marques matching maui mcgraw means megill methods minitrack model modern multi multidimensional naacl neff ofanlp ofinformation optimal over pages papadimitriou pods press principal printion probability probabliselephants proceedings processing raghavan readers references regression rehder research retrieval salton scale scaling schreiner sciagement science segmentation semantic seprocesses similarity simon societyfor sources special statistics story subspace summarization symposium system systems tamaki tent text topical trec understanding vempala visualizponents wolfe workshop york zhang http://doi.acm.org/10.1145/345508.345610 39 N e w Paradigms in Information Visualization able about achieved actual algorithm analysis anchors ankerst another apphcations approach approaches assa athina authors automatic baldonado bead best brenninkmeijer browsing chalmers chitson chose christian circle cluster clustering clusters coded cohen column columns computers conclusions conf contextual contributed contribution corresponding could create curse data design dimensional dimensionality display displaying document documents done down dragged drill earlier economou electronic engine epsi evolution exploration explorations exploring features filling fine first follow follows formed freely from fujitsu fulltext generate giving glance graphic graphics grinstein groups hemmje highlighted hoffman holistic ideal identified ieee improved include infocrystal information informationretrieval initial interactively intercluster interesting interests interface intl irrelevant irsg japan keim knowing knowledge korfhage kriegel kunkel labelling laboratories large learning less look lyberworld made main management manually mapping maps matrix measles menus milo more movement much multidimensional multivariate mumps naturally nature navigator neighbors nonlinear note number object order ormatzon over partially philip primarily primitive proc proe pull purpose queried query quickly radviz ranked reduction references refinement related relations relevance relevant repomtory representation restricting retrieval returned riiger rows rubella sammon screen search searching second seeding segments sensemaker sets setting sewraz shneiderman sigir space spoerry street structure study subject subset supported supporting technique thank that their them these this tool transactions transactwns tree used user using vacek vectors view visual visualisation visualization visualizations visualizatwn visually ways weed wilier winograd with word words work wsualization zervas http://doi.acm.org/10.1145/345508.345612 40 Latent Semantic I n d e x i n g M o d e l for B o o l e a n Q u e r y Formulation academia access advantages algebra alleviate allows also always american analysis asian aspects automatic berry better boolean brien buckley called characteristics cisi coffee combine communications computational computer conclusions cornell december deerwester diego dumais efficiency experimental extended figure first formulation formulations furnas generation handling harshman have however improvement indexing information intelligent international journal landauer language languages latent letsche levels lexical linear matching methods model models natural norm noticeable over park performance power precision problems proceedings properties proposed query quries recall references representation results retrieval review rocali salton science second semantic showed siam similar sinica smith society standard supercomputing syntactic taipei take than that theoretical thesis this traditional university using utilize vector with workshop http://doi.acm.org/10.1145/345508.345551 9 Evaluation of a Simple and Effective Music Information Retrieval M e t h o d access acoustic after alberta alexander alexandra algorithms alise allen also analysis annual applications approach artis assisted association august australia available baeza bainbridge barlow barry belknap benn berkeley bethesda brook brown bryce california cambridge canadian carolyn catalogue center century characters clare cliffs code collections comparison components computer computing conceiving concepts conceptual conference connectedness contour cunningham data database databases david denys department dictionary digital directory dissertation dlib document donna dowling downie duggan easie edition edmonton effective eleanor electronic englewood ernest evaluating evaluation experimentation faculty failures field fontes formation frakes from gerard gould graduate grams grove hall harman harold harvard henderson hewlett hill html http humanities index indexing information input interface international introduction inventory issues itworks january jean june justin kate keller kinnucan large libraries library lloyd lnformation london lutz machinery macmillan magazine management manuscript manuscripts mark mary maryland matching mcgill mcgraw mclane mcnab melbourne meldex melodic melodies melody memory menlo methods michael modern morgenstern munich murray musedata music musicae musical musicales musicians musicke musicology musifind national nelson notating ontario options ordinary organizations park parsons people phase philadelphia plaine pmgmatics poster prechelt prentice presented press procedures proceedings processing project psychological public publishers rabson rainer randel ranking references repertoire report representational representing research retrieval review revisited ricardo rism rmit rodger sadie sally salton saur scale science search secular selfridge series setfridge sigir similarity simple smith sources spencer stanley statistical stephen structures studies sutcliffe system systems tague technical technology text thematic themes theory tonta toronto towards trends tune tunes tuneserver typewriter typke uitenbogerd universe university unpublished verlag walter western william winkle with witten words wwwipd yasar yates york zealand zobel http://doi.acm.org/10.1145/345508.345539 2 Do Batch and User Evaluations Give the Same Results? across american anderror annual approximations aslib australia average baseline batch bell between buckley butterworths chin cleverdon comparing compressing computer computing conference cranfield data determining development diehl docrec document documents dublin each effective evaluation experiment exploring factors figures first forum from gigabytes group harman hersh human images improvement indexin information instance instances instrument interactive interest interface international jones journal keen lagergren length level library linear lndexing london managing matrix meadow measuring medicine melbroune mitra model moffat normalization norman nostrand number okapi over overview performance perspectives pittsburgh pivoted poisson precision probabilistic proceedings process project quarterly query recall references reinhold relationship relevance research retrieval robertson runs satisfaction saved science seen sigir similarity simple singhal sites societyfor some space sparck special swanson switzerland systems table text track trec trial user walker weighted with witten york zobel zurich http://doi.acm.org/10.1145/345508.345584 29 The Impact of Database Selection on Distributed Searching advanced agement algorithms allan american analysis applications approach april australasian automatic ballesteros based battle belkin broker byrd callan carnegie cluster colbining collection collections combining communication comparing computer conf conference connell craswell croft data database databases decision digital discovery dissemination distributed docu document does duxbury edition effective effectiveness effects effectweness emmitt ence engines ensuring evaluating evaluation evidence experiment experiments fifth first fourth french from fuhr fusion garcia generalizing gloss gravano gupta harman hawking hierarchies hull image inference information inquery institute intemet internet intl introduction isolated january johnson journal july june kantor koushik laird language large learnr lection libraries lnformation manretrieval mellon ments merging methods models modlin moffat molina multiple networked networks november over overview pages performance powell poweu press prey problem proc proceedings processing query ranking references report representation representations results retrieval rybalov sampling school science scipages search searches searching selection server shaw sigir sigmod sixth society source space statistical strategies swan syscollections system systems technical techniques technologies tems tenth testbed testing text theoretic third thistlewaite tomasic transactions trec university using vector viles visual vldb voorhees wide with yager yuwono zobel http://doi.acm.org/10.1145/345508.345578 26 Document Clustering using Word Clusters via the Information Bottleneck Method Noam Slonim and Naftali Tishby academic access adaptive advances agglomerative agnostic allerton almost analysis anderberg andrew annual applications approach arbitrary association automatic baker based bialek boostexter botflencek bottleneck brown browsing butterworths categorization category chert class classification cliffs cluster clustering collections columbus communication comparison computation computational conf conference constant corpus cover croft crtical cutting della demonstration desouza developments distributed distributional divergence document doeuments effects efficient eguchi elements empirical englewood english entropy etzioni expanded feasibility filter fine gather gother gram hall hard harper hearst hierarchic hofmann http hypothesis ieee incrementally indexing information interact interaction interactive iwayama john kachites karger label lang language large latent learning limit linguistics london machine management markovian mccallum measures mechkour mediated mediating meeting mercer method modeling models multi multiclass muresan natual netnews neural nips ohio page pages pedersen pereira pietra prentice preperation press probabilistic proc processing project projections queries ramsey recent reexamining references results retrieval review rijsbergen roussinov salton scatter schapire schutze science search semantic sequences sharmon sigir silverstein singer slonim smart sons statistical strategies study subsets system systems text theory thomas through time tishby tokunaga tolle tool toolkit transactions trends tukey using very webcluster wiley willett words yaniv york zamir http://doi.acm.org/10.1145/345508.345650 57 Stemming and its effects on TFIDF Ranking ablation about accuracy algorithm algorithms approximately based being better between binary cacm capable case codes comp computer correcting correction counting coverage cranfield damerau data dataset deletions demonstrated detailed detection development dictionary different dissimilar documents dokl each easy effect effective enhancement error errors eval evaluate evaluation experiments figure harman here hull impact impossible improve improvements inference insertions jasis kraalj krovetz length levels levenshtein ling lovins mech method morphology much number other others palce percent perf performance phys porter precision process program queries random ranged recall references regarding relevance resulting reversals rule seems sets short shown sigir similar sizes some spelling stemmer stemming stripping study such suffix suffixing technique than that these this topics trans used values very viewing vocabulary were while with words would http://doi.acm.org/10.1145/345508.345545 5 IR evaluation methods for retrieving highly relevant documents acadermc algorithmic allan american analytic annual approach approaches automatic background ballesteros belkin blair boolean borlund boston broglio callan case chamis cognitive combining communications comparative complexity computer conceptual conference conover croft databases department development devices dissertation document documentation effects effecuveness evaluation expansion experiments expression fifth filtering full gaithersburg green half harrnan hersh hickam iinen impact index indicators inen inference information ingwersen inquery institute interactive international introduction jarvelin john journal kantor keen kekalainen kluwer language libri life losee management manual maron massachusetts measures measuring medical method methodology models moffat national natural networks nonparametric online output performance position practical practice principle probabilist probabilistic proceedings processing publishers queries query rajashekar range ranked ranking references relationships relative relevance representatzons research retrieval retrieving rijsbergen robertson rvelin saracevic science searching seeking sigir smithson society sons sormunen ssertation standards statistics structure studies study survey syntagmatic system tampere technology term text textbook tnvison trec turtle university voorhees wiley wilkinson willett with wtde york zobei zobel http://doi.acm.org/10.1145/345508.345548 7 Event Tracking based on D o m a i n D e p e n d e n c y abstracts acknowledgements addition allan also annual apanese apple application apply approach approaches artificial association authors automatic average based beta better binomial boostexter boosting broadcast carbonell carp categorization comments company comparable complementing complexity conference content corpus creation darpa database decisions dependency detection development doddington domain earlier effective efficient electronic encoding especially european evaluate event existing expert exploring extending final foundation fourth from future gillick gomputational gonferenee grammar grant grishman handle hapter have ieige ilill improved included includes information intelligence international introduction japan joint jonference journal jsps kaneji lavrenko learning lexical lightweight like line linguistics literary literature lluman lowe luhn machine mandala mcgill mcgraw mechanized media method methods miller mixture model modern mron mulbregt multiple names network news newscasts newspapers ninth nouns okaxta only pages paper papka parser parsing performance pierce pilot plan precision press prmu probabilistic problem proceeding proceedings proceedmgs program promotion proper providing publishing rctmeval recall recognize referees references report reported research researchers researeh resnik result results retrieval retrospective robust roget sakamoto salton satoshi schapire science searching sekine semantic sigir similarity smoothing society some statistical stein stories stream strzalkowski study supported system tanaka task tateisi taxonomy technical techniques technology terminals tested text thank their thesauri this time tokunaga topic tracker tracking transcmplion transcmption transcription umass understanding understandtng university using valuable version watanabe which wise with wordnet work workshop would yamron yang york http://doi.acm.org/10.1145/345508.345621 43 The Use of Phrases from Query Texts in Information Retrieval addressed adequately analysis appear applications attached because benefits bucldey carballo cardle chunk complex component conclusion conference conveyed current document editor effectiveness ehunk eighth empirical expanding explore extraction fifth forth forum further future general guthrie have having honma htrul http important includes increase index information into investigation involving issues karlgren language leistemnlder list llow long lose management mano method mitra model modifiers more multiword narita natural noun obtained ogawa original perez phrasal phrase phrases post potential prepositional problem proc processing proe prohabilistic proximity queries real references report reported representation results retrieval riao second semantic should showed sigir significant singhal soflware statistical stop straszheim structuring strzalkowski strzalkowskl study syntactic terms text that three topics treatment trec using wang weighting when whole wilding words work world zhou http://doi.acm.org/10.1145/345508.345550 8 Improving t e x t categorization m e t h o d s for event tracking algorithms allan alvin amer amit annual applied approaches apte archibald assessment association automatic automatzc betrzeval boosting brian broadcast brown bruce buckley callan carbonell categorization classification classifiers cliffs coefficient cohen combination combining computing conference contextsensitive cosine croft curve damerau darpa david decisions detecting detection developm development different document doddington editor edward effective efficient englewood error eurospeech evaluation event events evidence examination expert experzments ezghteenth feedback filtering final first fiscus francisco from fzrst garofolo generalized george goets greece hall hampp human ieee imai improving informat information informatzon instance intelligent intellzgent international internatzonal jaime james jersay joeseph john johnson joon journal kamm kaufmann lafferty lain larkey lavrenko leah learning lewis liberman likelihood line linear machinery makhoul mark martin maximizing maximum methods michael mining model morgan multiple network news nguyen nineteenth nist nzneteenth oles ordowski output pages papka performance pierce pilot piscataway post prentice proceedings proceedzngs processing processzng properties przybocki publishers ralf rates recognitzon recognizer reduced reduction references relevance report research retmeval retrieval retrospective retrzeval retrzvval rhodes robert rocchio rover salton schapire schemes schultz schwartz sczences searches second segmentation shaw sigir signal singer singhal sista society soczety speech statistical study system systems task text thomas topic tracking training transcription transcrzption twenty understanding understandzng using victor volume voting walls weighted weighting weiss william word workshop yamron yang yield yiming yoram york http://doi.acm.org/10.1145/345508.345563 16 Automatic adaptation of Proper Noun Dictionaries through cooperation of machine learning and probabilistic methods acquisition advanced agency agichten alembic analysis applied august automatic automatically azzam based basili bikel borthwick brill case categories church coling colloeational communications computational computer conf conference corpora corpus cowie cucchiarelli cunningham darpa database december defense department description disambiguation discourse domainappropriate driven engineering english entity error evaluation evidence extension february finder finding finite from gaizauskas gale gazetteer generalizing generated george gnshman granada grishman harriman high humphreys inventory japan japanese journal july kaufinann kaufmann kyoto language large learning lexical linguistics machine marziali mateo mene message miller modelling models morgan name named nantes natural nmsu november nymble parser parsing part patterns pazienza performance phrase proc proceedings processing programs projects quantitative quinlan references research resources robinson roget rule schwartz science sekine selectional semantically sense sequences seventh shallow sheffield sixth spain specifications speech state statistical sterling study syntax system tagging technical texts trained transformation uncertainty understanding university used using velardi vilain washington weischedel word wordnet workshop yarowsky http://doi.acm.org/10.1145/345508.345538 1 Relevance Feedback with a Small Number of Relevance Judgements: Incremental Relevance Feedback vs. Document Clustering aalbersberg adding advantages allan annual applied biased boosting buckley bucldey butterworths cardie clustering conference development document effect efficient environment feedback filtering incremental information international learning length london mitra normalization optimization pages pivoted proceedings projections queries query references relevance research retrieval retrseval rijsbergen rocchio routing salton sanderson schapire schftze sigir silverstein singer singhal sixth smart summaries superconcepts text tombros trec using walz weights within zone http://doi.acm.org/10.1145/345508.345670 69 Integrated Search Tools for' Newspaper Digital Libraries access algorithms also analysis appear archives article automatic available bangalore based chandrinos comell computer conf content creation database decomposition dept digital document fagin from fuzzy gaithersburg gatford gatos gouraros hancockbeaulieu icdaw ijodl india information integrated intern international issue ithaca jones journal libraries library linking mantzaris multimedia newspaper nist okapi page paredaens parts perantonis principles proc queries recognition references report revisited riao robortson science seattle september singhal spring syrup systems technical term thesis tracking trec tsigris ttarman univ walker weighting http://doi.acm.org/10.1145/345508.345552 10 Phonetic Confusion Matrix Based Spoken Document Retrieval access accuracy algorithm american amir applications applying approach atlanta audio australia automated automatic auzanne based baseforms bayesian beaulieu blanchard broadcast broadcasts browsing cambridge carolina cbaivl cohen combining communications computer conference content cuevideo demonstration determination detroit develepment dharanipragada digitallibraries distributed document documents effect efficient expansion experiences experiments fast favero first foote forum franz from fung garofolo gatford hancock hauptmann hawaii head hicss hilton hindle home http icassp ieee ijcai image imperfect imperfectly independent index indexing indexng information intelligent international island james jansen johnson jones jourlin journal june kuhns laboratory lcassp learning learnspace lewis libraries lotus lunassen lund mail march maron maybury melbourne mercer model modification moore multimedia multiple munteanu networks news nist okapi open overview pereira petkovic philadelphia phonemic phonetic ponceleon prec precision probabilistic proceedings ptvceedings publication queries query radio real recognition recordings references relevance results retrieval retrieving robertson roukos scale sciences search searching second sept seventh seymore siegler sigir singhal sixth slattery society software sources south sparck sparckojones special speech speecld spoken spotting srinivasan stanford status strings study switzerland system tabs techniques terms text theoretic third time topic tprec track transcribed trec tuble uble university unrestricted user using video viswanathan vocabulary volume voorhees walker wechsler weighting witbrock with woodland word words workshop young zurich