http://www.informatik.uni-trier.de/~ley/db/conf/kdd/kdd2004.html KDD 2004 http://doi.acm.org/10.1145/1014052.1014082 28 Rapid Detection of Significant Spatial Clusters addison advances agrawal analysis applications approach arti automatic balakrishnan based biometry birch birkhauser calculations cant case charlton choudhary cial cient clark clustering clusters communications computational conf conference cpdc craft data databases deng design detect detection dimensional disease editor editors fast gehrke geographical geometry glaz goil grid gunopulus high inference information instance intelligence intl introduction investigation joint kulldorff lancet lange large learning leukemia machine mafia medicine method methods mgmt mining models moore multi multiresolution muntz nagarwalla nagesh nasca neill neural northwestern openshaw pages preparata press proc processing raghavan reading references report resolution samet saul scalable scan scholkopf sets shamos sigmod signi spatial springer statistic statistical statistics sting structures studies subspace systems technical theory thrun turnbull university verlag very waller wang wesley wiley yang york http://doi.acm.org/10.1145/1014052.1014094 40 On the Discovery of Significant Statistical Quantitative Rules aaai able accuracy actionable adams adjustment agrawal algorithm algorithms almost always analysis anomaly appendix applicable applying approach aronis artificial association associations assortment attribute attributes aumann based basket baskets bayesian beautiful because belief benjamini best better between betz beyond bolton bootstrap boston brijs brin broad buchanan butler canada cannot carcinogenicity case categorical chance change chapman chile clark clearwater cohen comparisons complex compute conclusion conference conforms considerations continue continuous contrast control controlling cooper correlations counting cross cubes data database databases dataset datasets datta decision decisions decrease decreasing desirable detecting detection determining diego differences dimensional direct discover discovered discoveries discovering discovery discoverydriven disease distribution draw drew driven dynamic each edbt edition edmonton efron eighteenth empirically estimate estimation estimator even example examples expect expectation explanation exploration exploratory extending false fast fifth figure fourteenth fourth francisco frank frawley from fukuda function further generalizing generate good groups guaranteed guide hall hand have hochberg hypotheses hypothesis icml identifying ieee implement implication increasing independent individual induction intelligence international introduction itemset iterative jensen john joint jose journal know knowledge knowledgebased kohavi large larger lead learning level levels lifetime lindell ling london lower machine management mani many market marketing mateo means median megiddo menlo method methods minimal mining model modeling models montreal moore morimoto morishita motwani much multiple national network newlands niblett ninth none nonparametric note nperm nsig number numeric oates olap once only optimized other outbreaks overly padmanabhan paper park pattern patterns pazzani perm permutation piatesky pods point possible powerful practical predict predictive presentation press problems procedure proceedings procs product property proportion provost pvalue quantitative quebec randomization rate rates real references relational relationships removed resampling research rodent royal rule rules santiago sarawagi scheme science search second selection sets seventh shapiro should showed sigkdd sigmod significance significant sigsqrules silverstein similarly sixth small smaller smallest society solution solutions some sons spain springer spurious srikant statistic statistical statistics still strategy strong study such swinnen tables technique techniques technology test testing tests than that their then theory there therefore this though through tibshirani tokuyama tool tools track traditional trees truly tsur tucson tuzhilin twentieth types ullman underlying unexpected unknown using valencia validation value values vanhoof variables very visualization wagner want washington webb well westfall wets when which whose wiley will with without witten wong words workshop world york young http://doi.acm.org/10.1145/1014052.1014074 20 Interestingness of Frequent Itemsets Using Bayesian Networks as Background Knowledge aaai about abraham achieved adding advances agrawal alexander algorithms also altos always american americanheart analysis analyze anand anatomy another apply approach approximate arti asia association associations attribute august automata autonomous background based basic bayardo bayes bayesian beautiful been belief better between bing blood bnintro books borders bottcher brief bucket building built cation causal chen cholesterol cial ciency cient ciently classi claus code collections computer computing conclusions conf conference contextual could course cubes darwiche daryl data databases david deal dechter decision department dethlefsen developed diego direction directions discovered discovering discovery distributions does dong downloadable driven dumouchel ectiveness edges eighth elimination emerging empirical encoded entropy erences especially european evaluation exception experience experimental exploiting factors fifth finn framework france frequent further future general generating gibbs given global grammercy graphical graphs gray guide guozhu hamilton hannu harinarayan heart heckerman heikki high highly hilderman hill however html http huang icalp imielinski implementing implication imporved impressions improvement increasing independence inference inform intel intensity interactivity interesting interestingness international intl introduction investigated issues item items itemsets iteratively jaraman jaroszewicz jective jensen jinyan journal judea july just kaufmann kevin know knowledge kodrato kumar lakshmanan languages large learning ledge levelwise library ligence ligent like lipids local loquium machine making malaga management mannila manually marginal math maximum mcgraw measure measures method methods microsoft might mined minimal minimum mining mitchell models modify morgan most multi murphy murphyk myllymaki nantes nding nement network networks noted novo obvious open optimized other paci package padmanabhan page pages pakdd paper partial patterns pearl pkdd poole practical pregibon presented press principle principles probabilistic probablisitic problems proc procedural procedure proceedings proceedinmgs programming property proved pruning publications ramamritham reasoning redmond redundant references regina reliable removing report represent represented research respect reviewer rewrite right rina risk rule rules sampling science score screening search selecting sets seventh shah shelf should sigkdd sigmod silander silberschatz similar simovici small solution some spain springer srivastava structure sudarshan suggested support surprising survey susanne suzuki swami systems taipei taiwan technical techniques that their theories third this tirri toivonen tool tools track trends ttcher tutorial tuzhilin uence ullman unexpected unifying university until update uronen used usefulness user using venky verlag washington william with workshop would wynne yiming york zaki zhang http://doi.acm.org/10.1145/1014052.1014121 65 Diagnosing Extrapolation: Tree-Based Density Estimation accuracy aids also annals anova approximation areas begin being black boosting both breiman cart cation cert ciently classi clean conclusions covariance create data demand demonstrated density detection device diagnosing diagnostic diagnostics dimensional distribution distributional dynamics economics elements environmental estimate estimated exact extending extrapolation forward friedman function functional gradient greedy harrison hastie have hedonic high hooker http implied inclusion inference information journal kaurmann learning likelihood machine management mateo measure mining mixture morgan natural olshen outlier point points prediction predictor prices probability problem products programs project provides quinlan references region regression reject relative report representation resistant resulting rubinfeld settings several shifts springer stability stanford statistics stone straight structures system technical that this tibshirani tool tools trees understanding uniform university unusual used variables variant versus wadsworth with york http://doi.acm.org/10.1145/1014052.1014091 37 The Complexity of Mining Maximal Frequent Itemsets and Maximal Frequent Patterns addison agarwal aggarwal agrawal algorithm algorithmica algorithms application apriori arikawa arimura asai association automatic ball based bayardo best boros burdick calimlim characteristics cient ciently closed closet company completeness complexity computational computer computers computing connected construction counting cuts data database databases depth discovering discovery distributions dmkd documentation dunham engineering fast feasible forest foundations freeman frequent from garey gehrke generating generation gouda graph graphs guide gunopulos gurvich hunt icalp icde icdm ieee informatsiya infrequent inokuchi interpretation intractability itemset itemsets johnson journal karypis kawasoe kedem khachiyan khardon know kuramochi kuznetsov large learning ledge line linguistics long machine mafia makino maniatty mannila marathe mathematical maximal maximum minimal mining most motoda nauchno ogihara papadimitriou parthasarathy patterns permanent pincer pkdd planar pods prasad probability problems provan radhakrishnan ramesh references regular retrospective rules saluja satamoto science search searching semi sentences sequences sequential seriya sets sharm siam sigmod spade sparse speci srikant stacs stearns strategies structured subgraph substructure substructures subtrees systems tekhnicheskaya that theoretical theory tkde tods toivonen transactional transactions trees ukkonen vadhan valiant vldb wang washio wesley xiao zaki http://doi.acm.org/10.1145/1014052.1014087 33 Probabilistic Author-Topic Models for Information Discovery allocation american analysis annual applied approach arizona arti attribution authorship based blei browsing buntine cial cluster collections computer computing conference cutting deerwester department development diederich dirichlet document dumais erten exploring furnas gather graph graphical harding harshman indexing information intel international jordan journal karger kindermann kobourov landauer large latent learning leopold ligence literature machine machines models operations paass pederson proceedings references report research retrieval scatter science semantic sigir society support technical temporal tukey university using vector visualization wampler with http://doi.acm.org/10.1145/1014052.1014135 79 Automatic Multimedia Cross-modal Correlation Discovery∗ albert algorithms analysis analytical anatomy annotation annual approaches asian attributes audio august authoritative automatic barabasi barnard based brin canada categorical cation christel classi clustering combinatorics comparison computer conf conference content cook creation cross cuts czerwinski databases decomposition deployment detecting diameter digital dimensional discrete dividing doyle dumais duygulu dynamic eccv eighty electric electricity engine environment erdos european external faces faloutsos february field fifteenth fifth first forsyth framework freitas from gong graphs hauptmann haveliwala human hyperlinked hypertextual icadl ieee ifip image images index indexing information instance intel interact interaction international january japan ject jects jeon jeong jordan journal july kamvar kanade kleinberg kluwer kolda large latent lavrenko learned learning leary lessons lexicon libraries library ligence ligent linguistic lovasz machine malik management manmatha march maron marsyas matching matrix media mining modeling models mori multi multimedia multiple nakamura name naming natural nature networks news normalized novel organized page pagerank pages pakdd palmer papadimitriou pattern paul personalizing pictures pods probabilistic proc proceedings quantizing raghavan random ratan recognition references relevance report research retrieval roach roussopoulos satoh scale scene search searching segmentation sellis semantic semantics semi semidiscrete sensitive sept seventh siam sigir similarity snell sound sources stanford statistical storage survey symposium systems taher takahashi tamaki technical terabyte tokyo tool topic toronto trans transactions transformation translation tree tzanetakis university using vector vempala video videocube videos vision vldb vocabulary volume wactlar walks wang wenyin wide with word words workshop world zhang http://doi.acm.org/10.1145/1014052.1014141 85 Generalizing the Notion of Support agrawal algebra algorithm algorithms already also alternative applications applied april apriori army association associations attri attributes based binary bollmann boolean bradley center cient comparison computer computing concepts conclusions constraints continuous data databases dawak demmel department describ details dimensions discovery discussion distribution dmkd erclique error existing explorations explore express extend extending fast fayyad february foundations framework frequent future general generalizing germany guntzer hafez have high hipp icde icdm ieee interest into investigate item itemsets january july kamb karypis kaufmann kumar limit linear many measures mining minneap minnesota model more morgan munich nakhaeizadeh nding nity nontraditional notion numerical ogihara olis omiecinski oolean ossibilities pages pattern patterns performance plan practical present prior publishers push raghavan reader refer references related research retrieval rule rules save science sdorra sequential sets several show showed shown siam side sigkdd skewed space srikant standard steinbach straightforward strong supp survey technical techniques that theoretical there this tkde tolerant university used utes vldb where with work xiong yang zaki http://doi.acm.org/10.1145/1014052.1014098 43 Predicting Customer Shopping Lists from Point-of-Sale Purchase Data abound adomavicius agrawal algorithm algorithms application applications applied architecture arena arti association assortment attributes bases bellamy blum brain brezin brijs build buying carlson case cation charles chile cial classes classi cohen communications company computer computing conference cumby customer data decisions department designing discovery editor ehrenberg factors facts fast foundations francisco frieze functions geyer government grocery hahsler hierarchy http human icml ieee improving incidence industry information intel interface international irrelevant issues jahn joint kannan kaufmann kellogg know language large learning ledge ligence limited linear littlestone london machine madison mccallum methods miami mining mitchell mobile model morgan natural neurocomputing newcomb noisy organization pages palm paper pashley perceptron perceptrons polynomial press probabilistic proc proceedings product programs psychological publishers purchase quickly quinlan recommender references repeat report reprinted retail review richards rosen rosenblatt rosenfeld roth rules santiago schulz science shavlik shrinkage snow srikant stasko storage study swinnen symbol symposium systems technical technologies text theory threshold time track tuzhilin uiuc uiucdcs usability using vanhoof vempala very webkdd wets when workshop http://doi.acm.org/10.1145/1014052.1014110 54 On Demand Classification of Data Streams aggarwal analysis attributes babu best birch cation changes changing cient class classi classify clustering clustream cock conference contains data databases datar dataset demand demonstrates diagnosing distribution domingos duda encer evaluate evolving figure framework gure hart high horizon hulten init intrusion issues large livny method mining models motwani network number oints pattern percentage pods processing ramakrishnan rate references result scene shows sigkdd sigmod size smallest stream streams synthetic systems test that time units used very vldb wang widom wiley with york zhang http://doi.acm.org/10.1145/1014052.1014137 81 Cluster-based Concept Invention for Statistical Relational Learning able accuracy accurate adding addition advances aggregation alexandrin algorithmic alternative analysis annals approaches attributes augmented author authorship background based better both built cameron cant citation citeseer class classes claudia cluster clustering clusterrelations clusters communities compared complex computing concept concepts conclusions conditional conference content contributed craven creation cument data database databases david dean debugging dels derive derived digital dimension dimensionality discover distance dmitry documents duction dzeroski ecting editors enter entities entropy ertext escul estimating existing experimental feature features finding flach flake foil foster framework friedman from future gary generate generation geto gideon giles gives graph grouped groups have hidden highly hill horvath icml identifying ieee included including increase induction inductive information interscience into intro introduction invention jones journal kaufman kirsten koller kramer lavrac lawrence leaps learning libraries link logic lyle machine mathias maximum mcgill mcgraw mean metho method middle mining mixtures models modern more muggleton nada only ontological oral original osal ositionalization over pages paper papers pavlov penno pennsylvania percentage perlich point predicate predicting prediction predictive presented presents press probabilistic proc process program programming programs prop provide provost publication quinlan raedt rather reduction references related relation relational relations relationships result resulted results retrieval richer rousseeuw salton saso schema schwartz scienti section shapiro shown signi slattery snowbird springer statistical statistically statistics stefan stephen steve streams systems tamas task tasks temp test than that their theory thesis those topics trends ungar unique university used using utah venue verlag were where which wiley with without word work wrob http://doi.acm.org/10.1145/1014052.1016917 103 ANN Quality Diagnostic Models for Packaging Manufacturing: An Industrial Data Mining Case Study absolute aceralia also among amsterdam analysis annual appears application applications applicatons approach arise arti atypical automation average based behaviour best between biase both bottom brussels calidad cases cation cause chapman check cial clementine clinton coating coils coincide comparing conclusion conclusions conference consistency control correlations could crisp cuadrado cuesta current data degree development diagn diagnosis diagnostic diez drawback drawn dusseldorf each ecsc electrotinning emerging enabling errors estimated european expected explains explotation factory failure ferrari figure finally forecast from fuzzy gonz gonzal gure heckenthaler hejlesen high hojalata holzknecht ieee ifac implementation improvement induction industry intel intelligent international internet japan just keplinger kerber khabaza ligence linear lines ling link lippo lisbon logic main management mars maximum mediante medium meeting methods milenium mills mining mode model modeling modelos models monitoring much musella namely narrower obeso observed obtained october only optimal optimisation optimizaci optimization organization originally other outliers outside oviedo pages para peregrina performance performances peters pittsburgh plant portugal previous problems proceedings proceso process processes product proposed provide quality range real references reinartz relative residues results rolling rule same selection sensors september series shown side siderurgico similar society speci stamatis standards steel stico study symposium system systems targeted technologies technology terms than that them these thesis thickness thicknesses this three time tinplate toba tokyo tools universidad validated values variable variables veri very view virtual visual were where which wide wirth with work workshop http://doi.acm.org/10.1145/1014052.1014123 67 SPIN: Mining Maximal Frequent Subgraphs from Graph Databases about accurately acknowledgement algorithm algorithms also annual applications approach apriori arimura asai bandyopadhyay based bases benchmark berhold biocomputing bioinformatics biokdd biology borgelt burdick bystroff calimlim carolina chemical ciency cient ciently classes classify closed closegraph coherrent compared compressing computational computing conclusion condensed conf conference contact current data database databases demanding deutsch discovery discussions dong eighth engineering equivalence european family fernandez ffsm finding forest fragments framework free freqeunt frequent from future garcia gehrke good gouda graph graphs group gspan gudes help helpful huan icde icdm ieee improvement indexing inokuchi international intl into investigate isomorphism itemset itemsets jack karypis kawasoe knowledge kuramochi labeled large least magnitude maps maximal mine mining molecular molecules molina motoda much muntz nding north number offers optimization order paci packing pages paper partially partitions pattern patterns performance pkdd practices presence present principles prins proc proceedings proposed protein proteins raghavan recomb references relevant report representing research residue rmed sakamoto scalability semi semistructured sets shao shen shimony sigkdd sigmod smaller snoeyink sparse speci spin state stored storing structure structured subgraph subgraphs substructure substructures such suciu symposium synthetic technical techniques thank this together transactional trees tropsha university vanetik very wang washington washio which will with work workshop yang zaki http://doi.acm.org/10.1145/1014052.1016921 107 Programming the K-means Clustering Algorithm in SQL adaptive aggarwal agrawal algorithm algorithms alternatives analysis appear association based berkeley binary birch bradley cation cereghini chaudhuri cient cikm classi clear clustering clusters conference data database databases decision dimensional discovery disk dmkd domeniconi dunemann dunn expression fast fayyad finding frem from gathering gene generalized graefe gunopulos harvey heytens high icdm ieee implications integrating jamil jected knowledge large livny locally lohman macqueen mathematical means method methods metrics mining multivariate observations omiecinski ordonez pages papadopoulos primitives probability proc queries ramakrishnan references reina relational robust rule sarawagi sattler scaling sets sigmod some spaces sqlem statistics stop streams symposium thomas tkde tree using very with workshop zhang http://doi.acm.org/10.1145/1014052.1014119 63 A Microeconomic Data Mining Problem: Customer-Oriented Catalog Segmentation actions added adopting agrawal algorithms alternative analyzed approximate approximating asodi association authority brijs cantly capturing catalog catalogs cient clas company compared completeness complexity computation computers conclusions considerations cover creating cross customer customers data discovered discovery ebdt erimental evaluation explorations feige formally formulation framework frameworks freeman from garey generalized generator goethals greedy guide have higher icdm interest intractability introduced investigated item johnson journal karypis kleinberg know knowledge kumar least ledge louie mannila manuscript maximal measuring methods microeconomic minimum mining model most mpis nite notion numb optimal optimization oriented overall pages pakdd papadimitriou paradigms particular patterns presented problem problems proc product profset programming promising raghavan randomized ranking real references relaxations retail rules safra segmentation selection selling semide several showed sigkdd signi software steinbach such supermarket swinnen symposium synthetic that theoretical theory this threshold unpublished using utility values vanhoof versions view wang wets with wong yield zhang zhou http://doi.acm.org/10.1145/1014052.1014083 29 Turning CARTwheels: An Alternating Algorithm for Mining Redescriptions acquisition agrawal algorithm algorithms american analysis approaches approximate association automatic bernstein bioinformatics biology botstein breiman brown cached cacm cactus carmel categorical cation causton cells changes chapman chromosomal cient cients classes classi cluster clustering complex computer concepts conceptual concise condition conjunctive construction cover data databases datasets decision discovery dissimilarity eisen elements ellman endent environmental erties euclidean expression fast figure fisher framework friedman from furnas ganti gasch gehrke gene generating genesis genetics genomic geometric gower grunstein halevy hall harel holstege human identifying ieee incremental inductive information intelligible international into jair janoski jennings john jones journal july kaufmann know knowledge koller lander landscap large learning ledge legendre ling logic machine management matching measures metric michalski microarray mined mining models module modules molecular moore morgan muggleton multiple nature networks nucleosome number olshen onse pages paper partitioning pereira perez pericliev pictures policy pottinger precision proceedings programming programs prop quackenbush quinlan rahm rainforest ramakrishnan record redescriptions redundant references regev regression regulators regulatory relevance research resp respectively rules schema science scienti segal shapira shore sigmod silencing similarity society sons srikant statistics stone storz studies sturn summaries survey systems their theoretical theory thomas through total track tree trees universal using valdes very vision vldb wiley with wyrick yeast young zaki http://doi.acm.org/10.1145/1014052.1014064 10 Fully Automatic Cross-Associations∗ access adomavicius agrawal algorithm algorithms also analysis annals appears approach approximately arithmetic asso association automatic automatica behavior being belonging besides better between bipartite birch blow boosting both bottleneck browsing candidate captures case cavegraphs caves chameleon ciations ciationsearch cient cikm cleary clickstream clock cluster clustering clusters coding colt column comm communities compress compression computer concept concepts conclusions conjunctive constraint cross crossasso crossassociationsearch cure dashed data databases datasets dayal decomposition decompositions deerwester dense description dhillon discov domingos dumais dynamic dynamics each ective elkan emerging epinions erent estimation explained expression extending fact faloutsos fast faster figure frequent friedman from fully furnas gene generalized generation graph graphs group groups guha hamerly handling harmonic harrison harshman have here hierarchical hinneburg hofmann homogeneous icml identifying ieee implementation increase indeed indexing inequality information inner integers irresp isit iswc iterations jasi jordan kamber karypis kaufmann keim kitsuregawa know kolda kontoyiannis kraft kumar landauer langdon large latent learning leary length linear linearly livny loop lossy mach madiman magic mallela management matlab matrices matrix means memory metho method methods microarray minimum mining mishra modeling modha montgomery moore morgan mosenzon most multi multilevel multimedia multiple multivariate nding neal nearly needs networks nips noise noiseless noisy numb number numbers observation observe ones only order organized overall overheads papadimitriou paper partitioning pattern patterns pelleg phenotype plot plots pods possible press princeton prior probabilistic problem proc properties proposal proposed raghavan ramakrishnan randomness rastogi reddy references regions regroup regularity relate research respect retrieval richardson rissanen rules satis sawtooth scalability scale scaled scales search seconds section semantic semidiscrete shim shortest shown shows sigir sigmod similar simultaneously size sizes slonim smal small solid some sparse spatial spectral spuriously srikant statistics step stopping structures successfully swaminathan swapping systems tamaki tang techniques text that their theoretic this three through thus time times tishby track transactions tree trend trends true trust tsdm tuzhilin underlying univ universal used using values variance vempala very vldb wall watts weiss were when wise with without witten words worlds zeros zhang http://doi.acm.org/10.1145/1014052.1014150 94 A Cross-Collection Mixture Model for Comparative Text Mining aaai across algorithm allocation baker barzilay between blei buhmann cation chakrabarti classi clustering communities computer concept conference correspondence coupled cross dagan data databases dempster detecting dirichlet discovery distributed distributional epinions extraction feldman from generic godbole hatzivassiloglou hearst hofmann http incomplete indexing international interspace jordan journal keyphrase klavans know knowledge laird latent learning ledge likelihood machine mappings marx maximum mccallum mckeown method mining multidocument mutual navigation pages principle probabilistic proceedings progress prospects references reformulation reinforcement research royal rubin sarawagi schatz semantic sentence shamir sigir sigkdd statist structural summarization text textual topics towards training untangling using words http://doi.acm.org/10.1145/1014052.1014100 45 Early Detection of Insider Trading in Option Markets acapulco across activity alberta analysis applications artificial august because behavior bernstein blaxton business changes clearwater conference data derivatives detecting detection diego discovering discovery early edmonton extracted fawcett fifth from futures goldberg hall hill hull information innovative insider intelligence interesting international investors john july kaufmann kirkland knowledge learning machine many mcmillan mcmillian mexico mining monitoring morgan multirelational nasd needed news noticing observation options other perlich plans potential prentice problem proc programs protecting provost publishers quinlan ranging references regulation relational securities shyr sigkdd solutions solve sonar sons spread summary system terrorist thakker trading uses westphal wiley workshop http://doi.acm.org/10.1145/1014052.1014148 92 2PXMiner - An Efficient Two Pass Mining of Frequent XML Query Patterns about acharya achieve agrawal algorithm algorithms arikawa arimura asai association both caching called chen cient ciently conclusion conference containment costly dasfaa data datasets dblp developed developing discovery equivalence exponential fast faster finally forest frequent from have impact important inclusion iner international investigate kawasoe large linearly miklau minimum mining number optimizations pass pattern patterns pods queries query reduce references response rules rundensteiner sakamoto scalability scales search semantic semistructured shows siam sigkdd sigmod since size space srikant ssplay structures substructure suciu support system tests than that this time times tree trees vldb wang xcache xing yang zaki http://doi.acm.org/10.1145/1014052.1014069 15 Systematic Data Selection to Mine Concept-Drifting Data Streams aaai able accuracy accurate active aggarwal algorithm also amount analysis approach approaches april arbitrarily argue arti august averaged babcock babu barbara based bases best better boston breiman build built california callaghan canada candidate cantly card cation change changes changing charity chen china choices choose choosing chosen chunk cial ciency cient claim class classi classify clustering compare compared complete computation compute computer concept conclusion conf conference consistencies consistently construct continually continuous credit cross crossvalidation cult data database datar dataset decision degree degrees demonstration detecting detection diagnosing different dimensional discovery discuss distribution domingos donation dong drift drifting drifts each engine ensemble error estimation evaluated evaluating even evolving example examples existing features first focs forests found foundations framework francisco fraud from function gain gambling given good greenwald guha happen have having helps high hong however huang hulten hyperplane hypothesis icdm ieee ignoring implementation implemented important indeed independent information insuf intelligence international issues jose july just justify keeps khanna knowing knowledge kong large learning like loss lower machine made madison management milshra mine mining model models more motawani motwani multi multiple national nineteenth nitely node observation once online optimal optimality oracle other outputs over pages paper particularly pattern physically pods points posterior press previous principles probability proc proceedings proposed quantile quantitative queries random randomly rate reading record references regression relatively remains right santa scale science second select sensible series several show siam sigkdd sigmod signi similarity situations size small software space speci speed spencer splitting stable statistics step still stream streaming streamminer streams street summaries symposium synthetic systems than that then there third this those three through time toronto training tree trees true under unselectively useful using validation various very vldb wang well when whether widom will wisconsin with without http://doi.acm.org/10.1145/1014052.1014084 30 Selection, Combination, and Evaluation of Effective Software Sensors for Detecting Abnormal Computer Usage abound abuse acknowledgments active agarwal also always anderson anomaly anonymous approaches april arising attributes based before behavior brodley building calls case challenges classifier collection comments company comparative complement computer computers concept conf critical darpa data defense details detect detecting detection development drift employees encouraging ertoz even excellent expected fahland final finally first focused forrest fort framework fully gathered ghosh have hill identification ieee information initiatives insider insightful internal intl intrusion intrusions irrelevant itap joshi known kumar lane latest lazarevic learning littlestone look lunt machine mcgraw michael mining misuse mitchell models monitoring network neumann nicely online opportunity ozgur patterns pearlmutter personal play pnrule present preventing privacy proc process profiles profiling program programming project provide quickly references report research responding reviewers role schatz schemes schwartzbard science security shavlik should siam skroch sole srivastava stolfo strategies strategy study such support surveillance survey symp symposium system systems technical techniques technologies thank that their them this threat understood undertake usenix user using volunteered warrender washington when will wish workshop http://doi.acm.org/10.1145/1014052.1014132 76 Sleeved CoClustering ∗ acad acknowledges acknowledgments agarwal aggarwal agrawal algorithm algorithms alizadeh also analysis application applications appreciation approximation automatic avraham best bibe biclustering biocomputing bipartite bittner carlo case cell chen cheng church click clustering clusters coclustering column columns computed computes computing conf conserved considered converges conversations coupled data decreases desired dhillon diliberto dimensional discrete disease distinct documents domany dougherty easily eisen employed enhanced enlightening erentiation except explicit express expression extracting fast fewer finding first follows form from function gehrke gene general generalized getz graph guan gunopulos helpful hematopoietic here high identi initial interpreting ismb iterative jected jective jones kasif large lazzeroni levine ling lymphoma maintains maps markers math matrix maxi melkman methods microarray midpoint mini minimum mining models mono monte moreover motifs murali natl nature norm operators optimal organizing owen paci pages park partitioning pattern patterns perceptive permute plaid pleasure presents problem proc procopiuc raghavan references residue reviewer rows scope seen self sequence sets several shamir sharan siam sigmod similarity sinica sleeve slonim solution some spaces special spectral squared starting statistica straus subsection subspace such symp tamayo that then theorem there tonical trent types using usually variables version wang width wish with wolf words yakhini yang http://doi.acm.org/10.1145/1014052.1014114 58 IncSpan: Incremental Mining of Sequential Patterns in Large Database ∗ agrawal association ayres bases bitmaps cheung cikm conf constraints data databases discovered discovering discovery dwarkadas engineering episodes expression flannick frequent garofalakis gehrke icde incremental information interactive july know large ledge maintenance management mannila march masseglia mining ogihara pages parthasarathy pattern patterns poncelet proc rastogi references regular rules sept sequence sequences sequential shim sigkdd spirit srikant technique teisseire toivonen update using verkamo very vldb with wong zaki http://doi.acm.org/10.1145/1014052.1016918 104 A System for Automated Mapping of Bill-of-Materials Part Numbers about agrawal approach artificial attempt august automatically automating baldridge based bierner bled califf chunking classification concluding conference damerau data describes description entropy extraction filtering first fledged full generalization hawaii hierarchical honolulu http ijcai information intelligence internal international items johnson journal june lafferty learn learning line machine mapping match maxent maximum mccallum methods mining modeling models mooney morton national nigam numbers opennlp ours package paper part pattern present proceedings references relational remarks research rules searching slovenia sourceforge srikant step stockholm such supplier sweden system text this uses using wide winnow with workshop world zhang http://doi.acm.org/10.1145/1014052.1014136 80 Estimating the Size of the Telephone Universe: A Bayesian Mark-Recapture Approach abundance agricultural agriculture alabama alaska american analysis animal arizona arkansas assessment association august available banding based basis bayesian biological biometrics biometrika borchers brooks buckland bureau calculating california capture carlo carolina chain chapman circular closed colorado commission communications competition computation conference connecticut cortes counts dakota danish data delaware demographic department developments diego directions discovery division dobra dynamics each early edition editors environmental estimates estimation extended extracting fast federal fienberg file fisher florida formatting fowler from future george georgia german gibbs gilks goudie hall hampshire hancock hawaii heterogeneity http idaho illinois immigration indiana industry international into iowa island jersey journal june kansas kentucky king know korn language large ledge leite level levene likelihood limfjord lincoln line london louisiana macmillan madigan maine mark markov maryland massachusetts maximum means methods mexico michigan milan mining minnesota mississippi missouri mixtures modeling models montana monte nebraska nevada nonparametric norris north ohio oklahoma oregon orted osterior oxford pages parameters part past pennsylvania petersen photo plaice pledger point pollock poole population populations poster poulovassilis practice pregibon present press probability proceedings recapture references related removal report research returns reys rhode richardson robert rodrigues rogers safe sampling seber service sfio sigkdd signatures silva simulation sixth size smith south spiegelhalter springer state station statistical statistics stats streams string summer survival table technology telephone tennessee texas theory these thousands total track trends uncatchable under university usage usenix using utah verlag vermont virginia washington waterfowl west wide wildlife wireline wisconsin with world wyoming yearly york young http://doi.acm.org/10.1145/1014052.1014134 78 A Quickstart in Frequent Structure Mining Can Make a Difference according acknowledgements active activity advances algorithm algorithms allow also analysis application applied arimura asai assumption atom available based berthold between borgelt broad cactus cancer canonical cantly chemical cient ciently classes classi closed closegraph code comp complete compounds computing conference constant data database dataset datasets dehasp discard discovering discovery discussing divided download edding enterprise entire enumeration erences erent erformed eriment eriments erties especially etter exploit features figure finding first forest forms fragment fragments free frequent from gaston graph graphs grateful have here hofer html http huan hybridtreeminer icdm idea ijcai inactive information inokuchi institute intel interested international into ipsj isomorphism karypis king known kramer kuramo kuramochi kyushuu large learning level ligent lists machine main management mckay memory michihiro mining modi mohammed molecular molecule molecules moto muntz nakano national ncidb nding nijssen numb number obtain obtained only osed osition oted ounds output pages part patterns practical presence prins proceedings processors prop providing raedt range references release report representative required results ruckert scale scienti sequences server sets seventeenth sigkdd signi signotes similar simple some source space spaces special ssdbm statistical structural subgraph subgraphs substructures summarized supp symposium technical test thank that their this thus time toivonen topic total track tree trees ulrich university unordered used using version wang washio which wildcards wise wish with workshop xifeng yang zaki http://doi.acm.org/10.1145/1014052.1014056 2 An Iterative Method for Multi-class Cost-sensitive Learning acknowledgments adaboost adacost addressed adjustment advances algorithms allwein also alternative another application approach approaches archive arti associazione bagging barlett based baxter binary blake boosting bradford breiman brodley brunk california called cation challenge chan choice ciale class classi coil comparing comparison computer conference convegno corvallis cost costs costsensitivity data databases decision decisions dence department descent direction discovery discussion discussions distributions domingos drummond ears eats eight eing elkan employ endent ensemble etween european example existing expected extensions fifth finally fourth frean freund friedman from fruitful fumera function future gbse geib general generalization gradient group holte however html http ieee imbalance incorp information insensitive intel interesting international investigate irvine issue italiana journal june kaufmann knoll know kohavi kunz langford learning ledge lessons ligenza line loss machine machines magical making margin margineantu mason mateo merz metacost method methods mining misclassi mlearn mlrepository modi modifying morgan multi multiclass nakhaeizadeh neural noisy note notes olsen onential oosting orating oregon ortant ortionate ository ossible other over pages paper perceptron predictors present press probabilities problems proceedings processing programs prop provides pruning quinlan rational reasonable reducing reduction reductions references regression related relationship relationships research restricted roli rosset saharon sampling scalable schapire science sciences sensitive sensitivity seventh singer sixteenth state stochastic stolfo stone straightforward studying such suggested supp system systems tausend terms thank that theoretic these thesis thinking third this time topic topics toward track trees twentieth under uniform unifying university unknown using vector viewed viii wadsworth weighting wellunderstood when with workshop wysotzki zadrozny zhang http://doi.acm.org/10.1145/1014052.1014152 96 A DEA Approach for Model Combination academic additional algorithm algorithms also amidale analysis annoted applications approach areas artificial bagging banker barber based bauer benchmark bibliography blake boosting broader california chanes class classification classifier classifiers clemen coelli combination combine combining comparison comprehensive computer conclusions conference contributes convex cooper cost data database datasets deabased deap department descriptions detailed develop dietterich dimensional dobkin domains dordrecht empirical england ensemble envelopment environment error estimating experiments extensions fawcett forecasting forecasts frank from further future given guide huhdanpaa hull hulls icml identify ieee implementations imprecise inefficiencies information intelligence international investigate java jongwoo journal kaoru kaufman kclass kluwer kohavi lane lawrence learned learning literature location machine management mathematical merz method methods mining model models morgan multi multiple netherlands note opitz optimal other outperforms oxford paper particular pazzani plan popular position practical present presented problems proceedings program proved provost publishers quickhull reduction references regression repository research review robust same scale schemes science seiford sensitive software solver some space srinivasan stanford study suggest systems technical techniques text that this through tone tools transactions univ university using value variants version voting weighting which with witten work workshop http://doi.acm.org/10.1145/1014052.1014089 35 Incorporating Prior Knowledge with Weighted Margin ∗ Support Vector Machines academic advances based bennett boston buckley categorization cation chang chemnitz classi classify collection computation conf conference data dellec demiriz ecml editors eled ervised european evaluation features francisco fung golub heidelb hersh hickam hopkins inference information institute interactive international joachims johns kaufmann kluwer knowledge large learning leone library libsvm loan machine machines mangasarian many matrix methods mining morgan neural numb ohsumed optimization pages press proc proceedings processing publishers references relevant report research retrieval rouveirol semi shavlik software springer supp support systems technical test text transductive univ unlab using vector verlag version with http://doi.acm.org/10.1145/1014052.1016923 109 Learning a Complex Metabolomic Dataset using Random Forests and Support Vector Machines academy achieved active also among analysis apparently applications applied apply available beecher belmont binary biol biological biotechnology bozeman breiman callahan cambridge carb cases cation cations choose classi classical classify complex computational conclusions cross cult current data decomp degree describ disease divide ective erform erformance erior error ersp explore feature fernie firmplus forests friedman from fung gaussian ghosh golden groups hand hawkins helix higher http ifcs illustrated innovations institute interactions into kernel kernels kristal learning leave linear lpsvm machine machines mangasarian matrices matson measurements metab method methods microarray moreover multiple national newest newton nitrogen nonlinear observed obtained olism olites olomic olomics olshen olynomial omics only opinion optimal optimization ortant osition other over package pages parameters passive pharmaceutical press proceedings random rate rates recent redox references regression report robust samples scholkopf science sciences second select selection several simmons simultaneous singular smola statistical stitt stone studies study supp support table technical technology tendency that these this tool trees truong tuning used using validation value vector vigneau wadsworth will with young http://doi.acm.org/10.1145/1014052.1014085 31 A Bayesian Network Framework for Reject Inference academic accurate advanced advances allison amemiya application banasik based bayesian berkeley bias biometrika california cambridge cation causal central centre clark classi cobb committee computing conference correction cost credit critique crook crossley data defeasible developments discovery does dordrecht dynamic econometrica econometrics economic ects editors eighth elkan ensity eople erformance error estimates evaluating evaluations example feelders foundations freund gabbay graphical handbook harvard heckerman heckman http icdm ieee ijcai imprecision improve inference information institute international into introduction journal kluwer know langford learning ledge machine management massachusetts mbrejinf microsoft mining missing model models multiclass murphy networks neural oaks observational ortionate overview pages pearl prediction press probabilistic probability proceedings processing prop publications publishers puhani quanti query really reasoning recent record references reject representation research role rosenbaum rubin sage sample sciences score scores scoring selection sensitive series seung shamir smets studies surveys systems technical thesis thousand tishby transforming tutorial uncertainty under university utrecht volume weighting with working zadrozny http://doi.acm.org/10.1145/1014052.1014124 68 On Detecting Space-Time Clusters aarts abbreviated accomplished adding addison address agrawal alamos alarms algorithm algorithms american analysis annealing appear applications applied approach appropriate arbitrary assuncao athas automatic balakrishnan behavior below biometry birkhauser both brain calculations cancer cannot cant carlo carnegie characteristics characterize close cluster clustering clusters combined communications company comparable compared computational conclusion conference confounding conjecture convergence correspond could covariates cylindrical dashed data detect detected detecting detection determine developed dimensional disease display division does dominant duczmal each earlier easily ective edited erent eriments estimated evaluating exhaustive exhibits exible experiment experiments explain extended extending factors fast feuer figure fleiss formally formulation framework from functions further gehrke genetic geometric geometrical geometry gives glaz goldberg group growth guarantee gunopulos handle happen have health heuristic high http improvements included indication information institute international introduction iterations john journal june kulldor laarhoven learning library likely lines machine make management many march mellon method methods mexico miller mining model modeled models modules monte moore more most movement multi national nding needed neill nips novel only optimal optimization other over overdensities pages parts phenomenon plot practical presented prevent prevention proceedings processing proportions public publishing pyramid quite raghavan randomization randomized range rates real references reidel report representing requires research resolution respectively results robust rudin satscan scan search services shape shaped shapes shrinkage sigmod signi simpler simulated since software solution solutions some something sons space spacetime spatial spread square statistic statistical statistics strategy stronger subroutine subspace such systematic systems technical that theory there this threshold thresholds time true university unlikely useful using value version very visualization weakness wesley when while wiley wilson with work http://doi.acm.org/10.1145/1014052.1014102 47 Eigenspace-based Anomaly Detection in Computer Systems agarwal analysis anomalies applied autonomic banerjee barford based baselining berman bipartite classics clustering communications communities components computing conference corbi data dawning dependencies detection determination dhillon ding directional disconnected discovering discovery distributed documents dynamic enterprise environments faults ganek generative ghosh graph gupta hopcroft ieee ifip international internet journal khan kline know kulis large ledge linked management mathematical mathematics matrices measurment method mining model natural nearly neogi network networks ninth nonnegative online operations pages partitioning plemmons plonka problem proceedings references sciences second selman separate seventh siam sigcomm sigkdd signal spectral systems using volume words workshop http://doi.acm.org/10.1145/1014052.1014128 72 Learning Spatially Variant Dissimilarity (SVaD) Measures afss algorithm algorithms alternating analysis antonio attribute based become before bezdek calcutta california categorization cation centroids cluster clustering comparison concept conference considered corresponding covariance crisp cscad data decompositions dempster dependent derive described dhillon diday diego discrimination discussed documents easily editor equation equations estimate euclidean every fair feature frigui from function fuzzieee fuzzy gini given gustafson hathaway hawaii here heuristic honolulu identi ieee incomplete international iteration january jdgk ject jective journal kessel keywords laird large learning likelihood machine make matrix maximum measure minimize modha nasraoui negative normalizing note notes optimization pages pattern problem proccedings proceedings quite recognition refer references regularization remain royal rubin same scad section shown similar simon simultaneous since society solved some sparse springer statistical systems text that this update used uses using value values verlag version weighted weighting weights when with would zero http://doi.acm.org/10.1145/1014052.1014115 59 Parallel Computation of High Dimensional Robust Correlation and Covariance Matrices algorithm alqallaf also american annals approach association asymptotically bias biometrics calculating cation computation compute computer conclusion correlation covariance data dataset department detection determinant dimensional disp driessen engineering ersion estimates estimator estimators experiments expression fashion fast feasible forum geijn genes gnanadesikan high implemented interface journal kettenring konis large levels make maronna martin matrices message methods minimum mining more multiresp multivariate only onse ositive outlier pages paper parallel passing performed plapack possible press proceedings random real references representing residuals robust rousseeuw scalable scale scatter science scienti series sets seventh shown sigkdd solve standard statistical statistics technical technometrics tennessee that them these this timely university using variables were when with zamar http://doi.acm.org/10.1145/1014052.1014072 18 Cyclic Pattern Kernels for Predictive Graph Mining∗ aaai accuracy advances against agrawal algorithm algorithms alternatives annual approaches arimura asai association automated backtrack based bases bayesian becker berlin berthold bioinformatics borgelt boser bounds burges california cambridge cance cant case cation chapter chemical christianini cient ciently classi classifying collins combinatorics comp comparing complexity computational computer computing conference considered convolution cortes cruz cycle cycles data database department depth deshpande diestel dietterich discovering discovery discrete driessens ebbinghaus ectives edition editor editors eigenvalue eighth electronic eled enumeration environments erations erics estimation etween fast fawcett fayyad feature finding finite flach flum forest fragments frequent from fuchs furer gaussian geometric ghahramani graep graph graphs guyon hardness haussler helma holloway however ieee imprecise induction inductive inexact information inokuchi international isomorphism jagopalan joachims journal karypis kashima kaufmann kernel kernels know kohavi kramer kriegel kuang kumar kuramochi language large learning ledge leen leslie level linear listing lkopf lnai lodhi logic london machine machines making mannila margin marginalized matching mathematical methods minimum mining model modeling mohri molecular molecules morgan multiplicity nakano natural nature networks neural nips nite nonvectorial numb onential optimal ounded ounds pages paper paths pattern persp piatetsky planar positive practical press principles probability problems proc proceedings processes processing programming provost raedt raghavan ramon rational read references reinforcement relational relevant reliability research results robust royal rtner rules santa scale scholkopf science search shapiro shawe showed siam sigkdd signi simple sivakumar smola smyth society spanning springer srikant statistical string stronger structure structures subgraph substructures support symposium systems tables tarjan taylor technical testing text than that theory thesis toivonen tomkins track training trees tresp tsuda union university unordered unreal upfal using usual usually uthurusamy valiant vapnik vector verkamo verlag vismara volume watkins with without workshop wrob zaki http://doi.acm.org/10.1145/1014052.1016922 108 Document Preprocessing For Naive Bayes Classification and Clustering with Mixture of Multinomials aaai acknowledgements advances algorithm analysis analytic annals assumptions basford bayes bayesian berger berkhin brunk cadez categorization cation chudova classes classi clustering coding comparison conference correcting csiszar darroch darya data decomp dekker discovery discussions dmitry domingos duda error european event explain generalized geometric hart icml icpr igor ijcai improve information international interpretation iterative john kamal karger know learning ledge like loss ltering machine machines many marcel mathematical mccallum mclachlan michael mining mixture models naive neural nigam oosting optimality osition output padhraic pages pattern pavel pavlov pazzani pedro platt proceedings processing productive ratcli recognition references rennie rish scalable scaling scene shih simple smyth sons sparseness squashing statistics supp systems tackling teevan text thank towards training under using vector vilalta wiley workshop would york zero http://doi.acm.org/10.1145/1014052.1016916 102 A General Approach to Incorporate Data Quality Matrices into Data Mining Algorithms according accordingly allocating although aminian approach approaches appropriate artificial assessing bagging ballou becoming being belief berry beyond bias breiman building builds business capturing central collection communications concept conclusion conference consider context control correction costs current data davidson decomposition detection determine documentable drift efforts embody english enhancement enhancing ensemble environments eqpd errors flaws from functions generated gone growing have hence hidden importance improve improving increasing information intelligence intensive international involve issue itself john journal kauffman kaufmann kohavi kubat learning lends limit linoff loss machine management manufacturing many mastering mathematics matrices matrix methodology methods mining model modeling models morgan naturally nature network organizations pazer perturbations perturbed pierce plus predictors preparation presence procedures proceedings process product profits propose pyle quality records reducing references replicate require resources science simple sons stochastic such symposium systemic systems tayi that them theorem therefore these they thirteenth this training tries typically undetectable useful using variability variance variety wang warehouse widmer wiley with wolpert work zero http://doi.acm.org/10.1145/1014052.1014125 69 Why Collective Inference Improves Relational Classification artificial bias chakrabarti classifi classification conference customers data decomposition discovery domingos dzeroski enhanced friedman getoor hyper hypertext indyk intelligence international knowledge koller lavrac learning link links loss mining models national network pfeffer probabilistic proc references relational richardson segal sigmod springer squared structure taskar text unified using value variance verlag zero http://doi.acm.org/10.1145/1014052.1014088 34 Scalable Mining of Large Disk-based Graph Databases aaai acknowledgements adopting alberta algorithm algorithms among analytical answering applications approach apriori based bayada berthold borgelt canada charles chemical cient ciently city closed closegraph common computer computing conclusions concrete conf conference constraints cook cormen data database databases december developed devising discovery disk djoko ective edition edmonton european examine example executable extensive finding forest fragment fragments france frequent from furthermore future general graph graphs grateful gspan gudes hill holder icdm important improve improving index indexing industrial information inokuchi interesting international introduction japan jiawei johnson jose july june karypis kindly know kuramochi large largest ledge leiserson like lyon maebashi management mcgraw methods mine mining molecular molecules motoda multiple other pages paper paris pattern patterns pkdd press principle problem proc proceedings promptly propose providing questions recognition references relevant research rivest ronald sasaki satoh scalability scalable sciences seattle second semistructured sept shimony sigkdd sigmod simpson stein structure structures study subdue subgraph substantially substructure substructures such support symp system takahashi taking terrsa thank this thomas trees values vanetik variety very washington washio with work workshop would xifeng zaki http://doi.acm.org/10.1145/1014052.1016914 100 Exploring the Community Structure of Newsgroups acyclic albert amer amos anal analysis anna appl application authoritative azar barabasi brin bringing cambridge christos chung citation collective communities computer computing conference czech czechoslovak daniel data david diameter digital dynamics eigenvectors element emergence engine engineering environment faloutsos faust fiat fiedler foundations frank from gibson graph graphs hisao horst hyperlinked hypermedia hypertext ieee indexing inferring internet jared ject jeev jeong journal june kannan karlin kleinberg latent lawrence library link liou manjara math mathematical matices matrices matrix mcsherry meshes meta michalis models motwani nature negative network networks newman nite order page pagerank pages papadimitriou parallel partitioning pennsylvania petros pittsburgh planar pothen power prabhakar press probabilistic problems proceedings processing property raghavan random ranking references relationships report saia santosh scaling science search semantic sergey shang siam sigcomm simon smallworld social society sources sparse spectral spielman stanford strogatz symmetric symposium systems tamaki technical technologies teng terry theory topology university unstructured vempala vinay wasserman watts wide winograd with works world yossi http://doi.acm.org/10.1145/1014052.1014117 61 Locating Secret Messages in Images accomp accurate activity against aggarwal algorithm algorithms allows also american amongst amount another applications applying approach arms attacks automatic available based baxter bearing berg binladen black capable capacity carefully carlo chain cipra cise color comparative computation concentrated concepts conclusion containing continual correctly could course cover covert curves data davidson decreases defeated defending defined detect detecting detection developing difficult digital dimensional discussion distributed downloads duan duric effectively embedded energized energy equations evenly farid features figure followed framework fridrich from function functions fundamentals goljan good graph grayscale greatly have hawkins hidden hiding high higher however html http icdm identified identifying ieee image images imperfections implies improve increases index inference information innocent innovative intensities interest internet introduction ising jain jajodia jjtc johnson july kamber kambhamettu kauffman kelley knorr knowledge level list location machines many markov mathematical media message messages method mining model monitoring monitors monotonicity monte monthly more morgan most neal networks neural news notion number order outlier outliers palette panacea particular paul photonics pictures pixels potential practical precision probabilistic probability proc processing properties provos putzmann quite race reasons recall reduced references regions replicator require restoration results robust same scans searching security send sending shekhar shown sigmod sips small software sorted spatial spatially spie starr state statistical statistics steganalysis steganographic steganography stego stegoarchive stegobearing study such support symposium system systems table taken technique techniques that then therefore they this though through throughout today toronto traffic transmissions typically unified uploads usatoday useful usenix using utley variance vector very visual watermarking website webterror west westfeld while white williams workshop world would http://doi.acm.org/10.1145/1014052.1014140 84 Dense Itemsets aaai accuracy acknowledgments advances agrawal algorithms also analysis another appear application applications applying approach approximation argued aristides association assortment attention bastide behavior believe between binary bingham blockeel boolean boulicaut bounding bradley brijs bringing budgeted bykowski calders case cation ceur challenges chapter cient classi closed collections combinatorial complement computer concept conclusion counts course cover coverage covering curves data databases datasets decisions demonstrated dense density department derivable developed diagnostic dimensions discovering discovery discussions donating duda editors elomaa enlightening error experimental external factorization fast fault fayyad fimi finland first florida formulae free frequency frequent functions further gamberger gave geerts generalization generalizations giannotti gionis given giving goethals greedy hand handling hart have helsinki high hollm http icdm imielinski implementations importantly include independent inductive inen information interest interesting international intersection introduced investigating issues items itemset itemsets jaakko june kdid keim khuller kleinberg klemettinen know lakhal large lavra lead ledge letters lnai lncs ltering mannila manuscript many matrix maximum means measuring melbourne method methods mielik minimum mining model models mohammed moss most naor nding need negative neural nition only open ordering other pages papadimitriou paper pasquier pattern patterns piatetsky pkdd possibility practice preparation press problem problems proc processing product proposition queries query raedt raghavan real references related remain reminiscent reordering report research results retail rigotti robust rules scene science section segmentation select selection sepp sequential sets seung several shapiro show shows sigmod smyth some springer srikant study subclasses such support swami swets swinnen systems table taneli taouil technique techniques thank that these this tiles tilings todorovski toivonen tolerant topic topics topk transaction tung tzvetkov university usefulness using uthurusamy vanhoof various verkamo verlag volume wang well wets wiley with without work workshop yang york zaki http://doi.acm.org/10.1145/1014052.1014101 46 Mining Coherent Gene Clusters from Gene-Sample-Time Microarray Data acad alizadeh alon analysis annual array assessing biclustering bioinformatics biology black broad calculation cant cell change cheng church churchill class cluster clustering colon computational conclusions conf conference data detection discovery distinct eighth erge experiments expression february fifth fold from gene identi intel inter international ismb june kerr large ligent ling lymphoma microarray minimum molecular natl nature normal number oligonucleotide otstrapping patterns probed proc proceedings recomb references reliability replicate required revealed signi spots systems tissues tumor types http://doi.acm.org/10.1145/1014052.1014116 60 Belief State Approaches to Signaling Alarms in Surveillance Systems acting actually again algorithm algorithms analysis applications approach approaches approximated arti assumed assumption balakrishnan based bean bioterrorism birkhauser cantly carnegie case cassandra cations chastic cial communications corresp curves cusum data dataset detecting detection development disease diseases distribution distributions does domains early ears ecause elief emerging encer endency endent england erent erform erforms erration ertas etter fact false figure future geographic give glaz have health heavily hold hutwagner hygine improvement indep indicate infectious intel journal kaelbling kulldor ligence littman maloney many martin medicine mellon methods might mining monitoring much need normal observable observed obtained onding onse oratory orting other outbreak outbreaks outp over paritally patterns plain planning plot pomdp positives preparedness prevention pselect raub references regionwise resp results rogerson routine salmonella scan seeman shows signi similar skewed slutsker smaller spatial state statistic statistics successive surveillance survellience system systems table take than that theory these thesis this thompspn thresholding tillett time treadwell uenza units university urban uses using values wales when with wong work works worse http://doi.acm.org/10.1145/1014052.1014143 87 A Generative Probabilistic Approach to Visualizing Sets of Symbolic Sequences advances algorithm analysis arti assp bach barnes based bishop cadez cambridge categorisation cessing chains cial class clavier clustering combined computation computer conf conference data databases delling dels developments discovery discrete distributed duction dynamic early editor erament ered evidence evolution fitting from generative girolami heckerman heskes hidden hollm icann ieee image information intel internal international intro jordan journal juang keyb know lampinen latent learning ledge ligence ligent machine magazine mapping maps markov meek merz mining mixtures mozer murphy music nato navigation networks neural neurocomputing ninth nips oard ographic oral organising organizing ository pages pattern patterns petsche press probabilistic rabiner references ripley scand scia sciences self series simplicial simula site smyth spatial springer statistical streams svens system systems temp text theories topic trait transactions tresp user verlag visualisation visualise visualization webkdd well white williams with ypma http://doi.acm.org/10.1145/1014052.1014131 75 The IOC algorithm: Efficient Many-Class Non-parametric Classification for High-Dimensional Data aaai acceptance access advances advertisements albert alexander algorithm algorithms analysis anchors andrew annals approximate arti arya based bases bentley best blackard blott boundary bumptrees california candidature cation ciaccia cial cient classi climbing cohen cole committee computational conf conference constraint covertype data database databases david digital dimensional dimensions distances distribution editors eleventh empirical evaluations exact expected expo fanty feature finkel forest friedman function general geometry gionis goldstein gray hammersley hans hashing hauptmann hierarchy high hill hirsh http hypersphere ichi ieee illustrated index indyk inequality information informedia instance intelligence international internet irvine isolet jock jonathan jorg journal katayama kaufmann kibler kushmerick large learning letter letters library lippmann logarithmic machine mark matches mathematical method methods metric moody moore morgan motwani mount multimedia mutation nding nearest neighbor netanyahu neural nicholas nips nist nonparametric norio november olympic omohundro optimal pages patella performance preparata presentation press proc procedure proceedings processing project prototype proximity psychological quantitative queries random recognition references report roger sampling satisfying satoh schek search searching segmentation selection september shamos shin shot silverman similarity skalak slate software spaces spoken springer statistics stephen structure study supervised survive systems task tasks technical thesis this time ting touretzky trackorganized transactions trec tree trees triangle twelfth uhlmann uncertainty university using verlag very video vldb weber with workshops yanjun zezula http://doi.acm.org/10.1145/1014052.1014142 86 Ordering Patterns by Combining Opinions from Multiple Sources above absolute accurate addressing after aggregating aggregation algorithm algorithms almost among association avoiding based baseline because best bias boosting cagliari case class classification classifier cmar cohen combine combining compared computer conclude conclusions conf criteria data decision degree dependency dietterich difference discovery disparate disparity does effective effectively efficient eighth ensemble exception experiments expert feund finally first five framework freund from generalization hedge helps ieee improve improves information integrated integrating interestingness introduces italy iyer jensen journal knowledge kumar learning line lists machine mean measure median method methods michigan mining multiple neural neville observation opinions order ordering other ozgur paper pattern patterns perform performance preferences presents previous problem proc processing proposed provided pruning rank ranking reduce regression relational report results right rule rules schapire sciences second selecting siam sigkdd similar singer smaller sources srivastava state system systems table technical that their theoretic things this unaffected university unsup unsupervised values weighted when with without http://doi.acm.org/10.1145/1014052.1014104 49 Visually Mining and Monitoring Massive Time Series abnormal accords acids adaptive aggarwal ahajournals alberta alexa algorithmica algorithms amaral analysis anomalies apostolico appear applications approach approximate architecture archive artificial assessment automated based benchmarks berlin biological biology bock boston boulder calendar cambridge canada caraca cardle case cavtat chakrabarti challenge chang chavarrias chen cheshire chiu chronic circ circulation cluster clustering cohen complex componenets component computational computer conference consumption content cortex croatia crochemore czumaj dasgupta data database databases dataset della demonstration denver detection diego different dimensional dimensionality discover discovering discovery display distinctive dubrovnik durbin dynamic eamonn eddy edmonton effective efficiency efficient electricity electronic empirical engineering european evolution experiences exploration explorations eyes faloutsos fast finding forrest francisco from full future gabglio garcia gasjeniec generale germany given glass goldberger government graphics ground hausdorff have hepatitis high hochheiser html http huang huettel human ideas identifying ieee immunology implications improve index indexing industry information intelligence intelligent interaction interactive interpretable isaac issues ivanov japan jarominek jessica journal judgements july julyaugust june kasetty keim keogh knowledge krogh languages lankford large launch learning lecroq leipzig lengths level linear local locally lonardi lopez lynnes machine mack management manchester manolopulos mark matching mccarthy meaningless medical method mietus milan mining mitchison mobile models molecular molina monotony moody motifs moving muller multi multivariate national nature need neuroscience novelty nucleic oates object ohsaki online oria overview ozsu pages paper park patel pattern patterns pazzani peng perceiving perkins physiobank physiologic physionet physiotoolkit plandowski practice prefrontal prepared press previous principal principle principles probabilistic proceedings processing program progress properties proteins quality quan quantitative queries query quest random ranganathan recognition record reduction references reinert representation research resource retrieval robot rule rytter sato scale schbath schmill science scientific search searches segundo selow sense sequence sequences sequential series shahabi shneiderman sigkdd sigmod signals similar similarity space spaces speeding spirals stanley statistica statistical streaming stringmatching study subsequence subsequences support surprise surprising survey symbolic symposium system systems tanaka task taxonomy temporal that theoria thesis tian time timeseries tokyo tool towards track trajectories transactions tree trend tsdma tufte type uehara understanding university unusual using valente visual visualization visualizations visualizing visualysis viztree wang washington waterloo waterman wavelet weber website week white wiederhold wijk with words workshop yamaguchi yokoi yoon zhao http://doi.acm.org/10.1145/1014052.1014103 48 Effective Localized Regression for Damage Detection in Large Complex Mechanical Structures adaptive aiaa algorithm analysis applications approach arlington artificial asce asceahs asme assessment backpropagation based baxter berlin bicanoc boston braun breiman calculations carnegie caruana chen civil clustering complex computation computers conditions conference congress conjugate connection continuum cross damage data databases density detection dewolf dimensional direst discovery diverse dynamics edelman engineering ester estimation estimators explanation faster feature feedforward forum francisco garrett gdbscan ghaboussi gradient hagan hajela hierarchical identification ieee ikeuchi incomplete information integrity international interpretation intrator ivan journal kamath kanapady kluwer knowledge kriegel kumar lazarevic learn learning least lecture linear localized machine make marquardt materials mathematics melbourne mellon menhaj method minimization mining mitchell mobile modal model models motoda multiple multitask multivariate network networks neural nonconvex nondestructive notes numerical orthogonality oxford perception pina pittsburgh powell predicting prediction press proceedings references regression report representation responses reston riedmiller robot royal rprop sander sandhu santos science scientific selection selective sensitive sensitivity series shen siam soares society spatial springer squares statistical structural structure structures structuring suitable sullivan symbolic szewczyk tamma target task tasks technical tests theoretical thrun training transactions transfer tsou university using variables veloso verlag visual with workshop zhao http://doi.acm.org/10.1145/1014052.1014057 3 Approximating a Collection of Frequent Sets agrawal applications arun associations bart bases calders ciently city completeness computers computing conference data databases david derivable discovery engineering etween european fang feller freeman frequent garcia garey goethals guide hector iceb ieee imielinski international intractability introduction items itemsets jeev john johnson know large ledge management mining molina motwani narayanan pages patterns principles probability proceedings queries rakesh ramakrishnan references sequential sets shivakumar sigmod sons srikant swami theory tomasz toon ullman very wiley william york http://doi.acm.org/10.1145/1014052.1014079 25 Incremental Maintenance of Quotient Cube for Median agarwal aggregates aggregation agrawal alaoui algorithms analysis approximation barbara based bases beyer bombay bosworth bottom cambridge canada chaudhuri cient ciently computation computational concept conf conference cross cube cubes data databases davey deshpande discovery exploiting formal formation foundations franzke galois ganter garcia generalizing godin gray group gupta harinarayan iceberg implementing incremental india intel introduction issues jaraman june know labio lakshmanan large lattices layman ledge ligence maintaince maintenance management mathematical mining missaoui molina montreal multidimensional mumick naughton olap operator order pages pellow performance philadelphia pirahesh press priestley proc quasi quass quotient ramakrishnan record references reichart relational sarawagi semantic semantics sept sigmod sparse springer structure summarize summary totals trees ullman university venkatrao verlag very vldb warehouse widom wille yang zhao http://doi.acm.org/10.1145/1014052.1014097 42 TiVo: Making Show Recommendations Using a Distributed Collaborative Filtering Architecture aggarwal algorithms analysis approach approaches artificial autoclass automatic based bayesian billsus breese buckley budapest canny cheeseman classification collaborative conference data delios discovery empirical factor fayyad fifth filtering filters finland fourteenth graph hatches heckerman horting hungary icml implicit information intelligence international item kadie karypis knowledge konstan learning madison management mining morgankaufmann nichols pazzani piatetsky predictive press privacy proceedings processing rating recommendation references results retrieval riedl salton sarwar sigir sigkdd stutz tampere term text theoretic theory uncertainty weighting with wolf workshop http://doi.acm.org/10.1145/1014052.1014111 55 A Generalized Maximum Entropy Approach to Bregman Co-clustering and Matrix Approximation address algorithms alternative american analysis annals applicable applications apply approach approaches approximation areas association attempts austin axiomatic banerjee based because been berkeley best biclustering both bregman censor cheng church classes clustering clusters coclustering compaq constraints contingency cover csiszar cult data decomposition decompositions dhillon direct distance divergence divergences dyadic eachmovie elements entropy euclidean exist expression fact factorization focussed formulation formulations from function gene general generalized ghosh guan hartigan have hierarchical hofmann however http icmb icsi inappropriate inference information interest interpret interpretable interscience introduced invaluable inverse involves involving journal kldivergence learning least linear loss main mallela many matrices matrix maximum merugu microarray minimum mining model modha much nding necessary negative negativity nips occurrence often optimization oxford pages paper paral partitional press primarily problem problems proposed puzicha quite recent references related report requirement research residue results seung several shown since singular special squared squares statistical statistics studied such tables technical techniques text that theoretic theory there these they this thomas three time topic types university unsupervised used using utcs value various which widely wiley with work would years zenios http://doi.acm.org/10.1145/1014052.1014112 56 An Objective Evaluation Criterion for Clustering ability algorithm algorithms application appropriate clustering clusterings comparing comparison completely details exist external fowlkes impossible incapable index information internal jaccard jective literature making mallows measure measures measuring methods metric mirkin note often optimize performance practically prediction predictionrelated predictive quality rand references several since successful such supervised tasks that their thereby therein these tries typically unsupervised variation various http://doi.acm.org/10.1145/1014052.1014077 23 Towards Parameter-Free Data Mining accelerate accurate algorithm algorithms allison amaral analysis anguelov annual application applications approach approximation april archive arima association august austria automatica automation available avoid badger based benchmarks benedetto berlin best bioinformatics biological biology boston buena caglioti challenge chemistry chen chiu circulation classification classifiers clustering coil columbia comment comparing comparison complexity compression computation computational computers conf conference constraints content convergence current cybernetics dasgupta data database datasets decision deformable demonstration description detection diego dimacs dimensional discovery discrete distance domingos eamonn edgoose edition effective efficiency elkan empirical entropy estimates european evaluation experiments faloutsos farach fast fastmap fault fifteenth flexer florida folias forrest fourteenth francisco from fsttcs future gada gatlin gavrilov genome germany glass goldberger goodman gunopulos hadjieleftheriou hausdorff heuristic hirsh hori house html http icdm icml ideas ieee immunology implications improve index indexing induction indyk inequality inferring information intelligent international introduction issues ivanov jose joshuago journal june kalpakis kasetty kato kearney keeping keogh knowledge kolmogorov kwong lake language learned learning length lessons letters level living loewenstern lonardi loreto lower machine magical making management manuscript mark market markov matching meaningless means measure measurements measures meeting melbourne memory metric microsoft mietus minimum mining mitochondrial model modeling moody more motwahl multi multimedia multiple muthukrishnan natural need networks neural noordewier novelty online oriented pages past pattern peng perkins phylogeny physical physiobank physionet physiotoolkit pitfalls practice press principle proc proceedings process puttagunta queries quinlan rapid ratanamahatana rearrangements recomb recommended references report representation requirements research review rissanen riverside rivest robotics rules sahinalp salzberg savari scientific searches second segment selection sequence sequences series shahabi shepp shortest siam sigkdd sigmod significantly similarity smyth space springer ssdbm stanley statistical stern stock streaming subsequences support supporting surprise survey symbolic symp symposium systems technical templates temporal theory thinking tian time traditional tree trees trend triangle truppel tsdma university unpublished using verlag vista visualization vitanyi vlachos wang washington wavelet which whole with workshop wyner yairi yianilos zhang zhao zipping http://doi.acm.org/10.1145/1014052.1014133 77 Semantic Representation, Search and Mining of Multimedia Content access anchor approach approaches baeza balance barcelona based basis basu best between boat both buciu capturing cation caveats circuits classi computing concept concepts conclusions conf conference crucial databases dependencies descriptor detection discriminatory electronic essential explain exploring face factor favor feedback food found framework frequent frequently gaede generally generic given good graph gunther gupta have huang icip ieee image imaging important including indexing international intl investigated jain jose just keywords kotropoulos kozintsev large learning leveraging lexicon machines marroqu media methods metric model modeling models most multidimensional multimedia naphade natsev nature navarro note novel occur other outperformed pages particular performance performing perhaps pictures pita power practice proc processing provide queries query random references relevance relevant reliability retrieval robust rochester rock scalable search searching selection semantic semantics should similarity slightly smith spaces spain speci spie springer stability statistical storage success support surveys systems technology tend that their theory therefore these this topic topics transactions tseng used uses vapnik variety vector very video vimsys visual vldb ways well weymouth which with yates york http://doi.acm.org/10.1145/1014052.1014071 17 Discovering Complex Matchings across Web Query Interfaces: A Correlation Mining Approach abstract accessible according accurately across actor agrawal algorithm also alternative among analysis approach approaches appropriate association associations attribute attributes author automatic automatically based baskets batini before bergman bernstein best between beyond blaisdell boundary brightplanet brin brings brunk bulletin cally categorizing challenges champaign chang chaudhuri choose choosing cient classify cleaning clustering clustweb collect comine committee comparative compare complex composition computer computing concluding conf conference consider correct correlated correlation correlations count cross cult cupid current data database databases deep department dept design develop development different differentiation discussion disparate doan does domain domains domingos done each edbt effort empirical engineering engr evaluation expect experiments exploration explores first formal framework further fuzzy ganjam ganti generalizing generic given gravano halevy have help heterogenous hidden hope however http ieee illinois imielinski implications instance integration interest interesting interestingness interfaces introduction investigated investigation ipeirotis issues items itself journal july know knowledge kumar large last learning lehner lenzerini machine madhavan market martin match matching matchings mathematical measure measures metaquerier methodologies mining model more motwani movie name navathe observation observations observed occurrence often omiecinski online open opportunities other overlapping pages paper parsing particular patterns porter porterstemmer probe probing probings problem proc proceedings queries query rahm reconciling references relationships report repository results return reveal right robust rosenthal rules sahami same sample scale schema schemas science second selecting seligman semantic send sets several sigkdd sigmod silverstein smith solve some sources speci specifically srivastava star statistical statistics stemming still structured study such suited summary surfacing survey surveys swami syntax systematic tackle tartarus tech technical techniques that there these they third this threshold through time trans trial uiuc uiucdcs unclear understanding university urbana using validate value values vldb warrant well where whether which with workshop york zhang http://doi.acm.org/10.1145/1014052.1014090 36 Exploiting A Support-based Upper Bound of Pearson’s Correlation Coefficient for Efficiently Identifying Strongly Correlated Pairs acknowledgments addison addition agrawal algebraic algorithm almost analysis answer approach army association associations attributes auspices based baskets bayardo bayes behavior beyond brin bucila burdick calimlim called cambridge candidate categorical cations center change cient cities classi cohen complexity computation computational computer computing conceptual conclusions conditional conf constrained constraint constraints correlated correlation correlations cost cross data database databases datar demonstrated dense department designed dimensional dimensions directions discovery distribution distributions dual dualminer dumouchel ecology empirical engineering enumeration erent erformance eriment erimental eriments erty etween even examine execution exhibits exploratory extend eyond figure finally finding first free frequent fujiwara functions future gehrke generalize generalized generalizing generated generation gionis grahne grant gunopulos hide high human icde icdm ieee imielinski increase increased increases indicates indyk interesting introduction item items itemset itemsets january jermaine journal kachigan kifer know kumar lakshmanan large least ledge like linearly llnl lter maintained market massachusetts maximal measures methodology minimum mining minnesota model monotone monotonic motwani multi multivariate nasa negatively numb numeric optimized orted osed otential other ound pages pairs pang partially particular patterns performance plan playing pregib press principle principles prop provided pruning quantify queries query radius rank ranksupport rastogi ratio ratios real reasoning recall references report representation research result results reynolds rule rules rymon savings scale science screening search second seek seen sets several shekhar shim show shows sigkdd sigmod silverstein similar slightly standard statistical step supp swami synthetic systematic taper technical that their there this threshold thresholds through time tkde transaction transactional trend twin ullman under university usersp using wang wesley when which white will with without work xiong yang york zipf http://doi.acm.org/10.1145/1014052.1014092 38 GPCA: An Efficient Dimension Reduction Scheme for Image Compression and Retrieval abbadi achlioptas adaptive aggarwal agrawal aigrain algorithm algorithms america analysis applications approximate approximations baltimore bangalore barbara based california carlo castelli characterization cient clustering cognitive component computation computations conference content csvd data databases datasets decomposition dhillon dimensional dimensionality drineas dynamic ects edition eigenfaces estimation evaluating expression faces faloutsos fast fastmap focs framework frieze gene generalized golub graphs guan hierarchical high hopkins human icde icml ieee images indexing india johns jolli jose journal kannan kanth kirby large loan madison matrices matrix mcsherry media minimum mining monte multi multimedia nding nearest neighbor neuroscience number optical pages pentland petkovic pods press principal proc procedure proceedings rank ravi recognition reduction references representation residue retrieval review sampling santa search searches searching seconds sedighian selectivity shen siam sigmod similarity singh singular sirovich society soda space spie springer squared state stoc structures subspace third thomasian time tkde tools traditional transformed turk university value vempala verlag vinay visual visualization wisconsin york zhang zhou http://doi.acm.org/10.1145/1014052.1014095 41 Fast Mining of Spatial Collocations∗ advances agrawal algorithm algorithms applications applied association august based bases brinkho cation characterization chawla cient class classi clustering complex computational computing conf conference cure data database databases dbms density dent detection dewitt discovering discovery dynamical ective engineering erski ester fast frequent frommelt geographic geometry gouda guha guttman hall hand huang icdm ieee index information international introduction join joins know kriegel large ledge ling location mamoulis management merge method methods mining morimoto multiway munro neighb oring pages papadias partition patel patterns prentice preparata primitives proc processing rastogi references relationships results rules sander science searching seeger septemb sets shamos shekhar shim sigkdd sigmod skrikant spatial springer sstd stefanovic step structure summary supp symp symposium syst temporal threshold tour trans trees trend using verlag vertical very volume without xiong york zaki http://doi.acm.org/10.1145/1014052.1016913 99 Interactive Training of Advanced Classifiers for Mining Remote Sensing Image Archives addison advances agship aksoy analysis ancillary applications approach atkinson august automated base based bayesian binello breiman brooks browsing building canada cart cation change chen classi clinic cole communications complete computer cover data databases department digital duda editor elevation engineering extraction features february fischer franklin frequency friedman from frontiers fusion geoscience gonzales grammar grss haley haralick hart health http huang ieee image induction information intel intelligent interactive international introduction invariant jensen john july june kaufmann knowledge koperski land landsat langley lawrence learning levien ligence lines little machine manjunath marchisio mayo mcneely miller mining minnesota missing model modeling monitoring morgan multisource musavi nasa natara november october olshen pages partitioning pattern photogrammetric proceedings processing programs publishers quinlan recursive references regression remote remotely report research retrieval ridge robot rochester rogan rotation routines rpart rubin rule scene science scienti sensed sensing shapiro simon sons space statistical stone stork stow symposium systems technical techniques terra terrain texture therneau tilton toronto transactions tree trees tsatsoulis tusk using visimine vision visual volume wadsworth washington wesley wiley with woods workshop world wright http://doi.acm.org/10.1145/1014052.1014080 26 Mining, Indexing, and Querying Historical Spatiotemporal Data access advances aggregation agrawal algorithm algorithms allo approaches asso asynchronous based bases cation cessing chen ciation cient clusters comparison computing conference considine cyclic data database databases density developing discovering discovery dong doridis dynamic engineering erio erschatz ester event evolving extending fast frequent gunopulos guttman hadjieleftheriou hellerstein identifying ieee incremental index indexing indyk international interval ject jectories jects jensen journal know kollios koudas kriegel large learning ledge machine management massive metho mining mobile moving muthukrishnan noise novel oral ozden pages papadias partial partially patterns peng pfoser proc queries query ramaswamy references representative rules salzb sander schemes searching sequences series sets silb sketches spade spatial spatio spatiotemp srikant structure surveys symposium system technology temp temporal theo time timestamp transactions tree trees trends tsotras tsoukatos unknown user using very vldb wang with yang zaki http://doi.acm.org/10.1145/1014052.1014059 5 Recovering Latent Time-Series from their Observed Sums: Network Tomography with Particle Filters. aaai accuracy accurate active actual advances airoldi american annals approach association bayesian bejerano bell berzuini better bhattacharyya black breitbart bursty butterworths cald carlo carnegie case casella ciency colors communication computation computing contingency continuous count counts dashed data davis dels destination determination diagnosis diot directions discovery distributed distribution donoho doucet dynamic editor editors estimating estimation estimations evaluation example existing faloutsos fast fienb figure following freitas frequency from garofalakis geometry gilks gordon grabarnik gravey greenb gure harvard hidden higuchi iannone ieee inference infocom information informational intensities internet iterative jackson journal korn labs language large latent liang likeliho lines link loads lund manage mandelbrot markov massive mathematical matrices matrix measurment medina mellon metho methods mining monte moving multi network networks node noms notice observed odintsova ology organizing pages paper performance physical practice press princeton principle prints probabilistic probing problem proceedings pseudo rastogi real reconstructed reconstructs recovering references relative research rish roughan router royal salamatian scalable scale scienti second selected self sequential series several sigcomm sigkdd sigmetrics skewed society solid source spikes springer statistical statistics strategies structure studies subnet sums systems table taft target tebaldi technical techniques their theoretic theory this time tomography towsley track trade university using vanderb vardi varying vaton verlag versus viel wang well west workshop zhang zipf http://doi.acm.org/10.1145/1014052.1016915 101 Feature Selection in Scientific Applications academic accuracies accuracy acquisition adequate algorithm also always analysis applications approach astronomers astrophysical automated automatic bands based becker bibby breiman brodley browsing calculate cantu case cases catalog cation classi common comp computer conference considerable contain create creating crucial culties cybernetics danyluk data decomp demonstrated dimensionality dinstein discovery discussion diverse double ecause ectral editor editors eling endent engineering ensure erformed erforming erience eriments error example examples faint fairly feature features filters finding first fodor four francisco frequency frequently freund friedman from fusion galaxies generate generating goal greedy haralick helfand heuristics high higher huang human hybrid identify identifying ieee illustrate image imagery images implements improve indep indicate insights intel interest interface international into irrelevant ject jective jects journal kamath kaufmann kent know knowledge learning ledge level ligence like likely limits linear machine mallat manjunath manually many mardia mateo medians methods might mining mistakes more morgan multi multivariate nature need newsam noiseless noisy observations obtained olshen onents ones only oosting orientation ortant osition outliers pages pattern phenomena preprocessing presented press problem problems proceedings processed program prone provide providing publishers radio reasons reduction redundant references regression relevant remote removed removing representation resolution resources restricts result results retrieval saitta satellite save schapire science scienti scientists search selection sensing settlements several shanmugam signal simple simpler size smoothed sophisticated source spie statistics stone study success such suggests survey systems tang techniques technology textural texture that theory these they thirteenth this three time together tools training transactions trees twenty underlying used using valid volume wavelet well were which while white windows with work worth wrapp http://doi.acm.org/10.1145/1014052.1014130 74 A Framework for Ontology-Driven Subspace Clustering algorithm anders architecture arep based biclustering biology bipartite botstein brown bsed campbell cation cell cerevisiae cheng church clustering clusters comprehensive compression computer conference consortium cycle data databases density department determination dhillon dimensional discovering documents eisen entropy ester expression extraction fasicicles flexible futcher gene geneontology genes genetic genetics graph harvard high http hughes hybidization identi intelligent international jagadish jason kriegel large lyer madar matrix microaray microarry mining molecular nature network noise numerical ontology pages partitioning pattern proc references regulated report sacccharomyces sander science semantic sets sherlock sigkdd sigmod similarity spaces spatial spectral spellman subspace system systematic tavazoie technical tendency using vldb wang with words yang yeast zhang http://doi.acm.org/10.1145/1014052.1014120 64 k-TTP: A New Privacy Model for Large-Scale Distributed Environments ∗ according agrawal alberta algorithm algorithms anonymity association august based between bounds cambridge canada candidate candidates canu case changed chile clifton completeness composition computation computations concurrent condition controller correct counters counts countuv criterion crypto current data database didates discover distributed dmkd each edmonton electronic encrypted erences evaluated fast february fuzziness game generate gives goldreich have helger honest horizontally http icdm initiate jority journal july june kantarcioglu know last least ledge less leveraging lindell link lipmaa lncs lower main majority majorityrule massachusetts meaning mental micali mining model multi multiparty needs november numunc oblivious occasion october oded order pages paper partitioned partitions party peer pinkas play presented preserving privacy privacypreserving private proc procedure prot protecting protocols query references rule rules santiago scalable schuster second secure self september sigkdd society sources srikant stoc survey sweeney systems than that then theorem this uncertainty vaidya value values vertically vldb voting washington weizmann when whether whose wigderson will wisdom with workshop zero http://doi.acm.org/10.1145/1014052.1014081 27 Machine Learning for Online Query Relaxation aaai adding addison algorithm allow along also amer angeles answering answers applications approach approximate arti aspects attribute attributes baeza based bases between boolean brazil california cation cercone chakrabarti chaudhuri chen chiang chow cial clustering cobase college communications computer conceptual conference constraints continue controlled cooperative corella data database databases deductive department development directions discovery documents domain driven edbt editors empirically engineering error evaluating expert exploiting exploration explore expressing extend extending extensible failing fast february fifth figure finally first flex flexible framework from fuhr furtado gaasterland generalization godfrey graphics gravano grosjohann have huang ieee induced inference information informative initiative integrate intel intend intentional interactions interacts interested interface interfaces international introduced janas janeiro jority journal kanza kaplan kaufmann keim know kriegel language laptops large larson learned learning lecture ledge ligence ligent loqr machine maryland mechanism mehrotra merzbacher method minimization minock mixed model modern modi module more morgan motro multidimensional natural neto notes novel null october online ortega over pages paper park pattern performs pima plan porkaew possible preferences presupositions principles problem proceedings programs provide providing publishers quality queries query quinlan relax relaxation relaxations relaxes relevance research response responses results retrieval ribeiro robustness rules sagiv scalable science seave second selection semistructured seventh several shown sigact sigart sigmod sixth small society software srivastava subset successfully symposium system systems target technology that then theobald thesis this through tolerant towards track trade transactions tree type university used user using value values various vast verifying very visdb visualization water wave weikum wesley which wiederhold wisc with work workshop would xirql yahia yang yates yesil http://doi.acm.org/10.1145/1014052.1014138 82 Identifying Early Buyers from Purchase Data ability adopter agrawal algorithm algorithms approximation argonne arising bass behavior benson budap categories cial conf consumer customers dels determination domingos downloadable dsdp dual duct durables emans enson from growth http hungary implementing improved innovation innovators instead jagopalan jority june knowledge launch maha management marketing maximum mining muller national network networks newsgroup nite oratory package problems programming references research richardson satis scaling science semide sharing sites software srikant srivastava targeting tech unix using usion value viral when williamson worthwhile http://doi.acm.org/10.1145/1014052.1014073 19 Mining and Summarizing Customer Reviews aaai abstract abstracts access according acquisition adjective adjectives agrawal algorithm also amazon analysis answering apfa appear applied approaches associates association automatic balancing based beckwith believe bloedorn boards boguraev bourigault bruce buckley cambridge carbonell cardie case characterization chen church classification classifications classify coling colloquium combined combining companion complex computation computational computer condensation conference constructing content corpora customer cutting daille data database dave dealt decomposition dejong demonstration detecting detection development direction directions discourse discriminant discussion dissertation document documents domain down eacl editor effects electronic emnlp engineering english erlbaum especially european evaluation extracting extraction fast fastr features fellbaum filtering finn formal foundations from frump fukushima fuzzy gallery genre genres gold goldstein gradability graph gross hahn handbook hanks hara hatzivassiloglou have hearst html http huettner hypertext identification ijcai implementation indexing individu infogistics information instance integrating intelligent international interpretation introduction issue jacobs jacquemi jacquemin journal justeson kantrowitz karlgren katz kennedy kessler knowledge kupiec kushmerick laboratory language lawrence learning level lexical lexicography lexter limitations limsi line linguistic linguistics literature litman logic machine main management mani manning manual market matching mckeown message methods metrics might miller mining mitkov mitra mittal model modeling morinaga multi mutual natural need needs nlprocessor norms note notes nunberg operational operators opinion opinions orientation overview oxford pages paice pang paper parsing paul peanut pedersen pennock performed perspective point powerful practical predicting prediction press proc proceedings processing product promising pronoun properties prospects question quiet recognizing references refinement reimer report representations represents reputations research resolution retrieval reviews rule rules sack salience salton scalable schutze search segments selection semantic sentence sentences sentiment settings sigir simple singhal smyth software some sparck special spring srikant standard statistical stock strategies student study style subasic subjective subjectivity summarization summarizer summarizing summary symbolic symposium syntax synthesis system systems tagging tait tateishi technical techniques term terminological terminology tetreault text textanalysis texts that themes they three thumbs tong toolkit track tracking trainable transfer turney tutorial typing understand university unsupervised used using vaithyanathan very view vldb volume what wiebe wilson with word wordnet workshop yahoo yamanishi http://doi.acm.org/10.1145/1014052.1014058 4 Mining Reference Tables for Automatic Text Segmentation aaai access accurately acero adelberg agents algorithm algorithms annual applications approach arti aurora automatic automatically autonomous bilmes borkar boundary bradshaw bulletin califf cannot cation challenges chaudhuri chen cial cient classi cleaning cleansing code cohen collins combining conditional conference consumption crescenzi data database datamold deng deshmukh dictionaries dirty discovery discriminative documents droppo dynamic eacl editors elds embley emnlp empirical engineering entity entropy etzioni eurospeech evaluation experiments exploiting extracting extraction factorized freitag from fundamentals fuzzy ganjam ganti gazetteers goodman grishman grover hall have hernandez hidden hierarchical hmms iaai icml ieee induction information integration intelligence international into jiang juang kaufman knoblock knowledge labeling language lapata large learned learning lerman machine markov martin match maximum mccallum mecca meeting memory merge merialdo message method methods microsoft mikheev mining minton model modeling models moens mooney morgan motwani muller muslea named national natural nist nodose online optimization ordering pages pattern perceptron pereira prentice press probabilistic problem proceedings processes processing przybocki purge rabiner random real recognition record records references relational reliably report resulting roadrunner robust rohanimanesh rosenfeld rules sarawagi school scie seattle segmentation segmenting selected semi semistructured sentence sequence seymore sigkdd sigmod singer sites sixteenth size smarttagger smoothing speech splice springer stochastic stolfo structure structured structures structuring study summer sutton technical techniques text theory third tool towards training tutorial understanding unsupervised uweetr verlag vldb what with without workshop world wrapper http://doi.acm.org/10.1145/1014052.1014068 14 Fast Discovery of Connection Subgraphs about access accesses across actors actresses adding addition additional additive addressed adopt advances advantage after aggregate albert algorithm algorithms allow along also always america among analytics another answers appendix applications appropriate april architecture association asymmetric attributes august average barabasi based basic because been begin best better between bias biological branching brandes brief brin bringing build candidate cannot cantly captured captures careful case cases categorical cation cations cell chandra changes characterizes chavet chilean choice choose chooses cikm citation close closeness closer closest clustering code coetzee columns common communities community commute comparably compares comparing comparison comparisons completely complex computation computational compute computer conclusions condition conditions conference connected connectedness connection connections connectivity consider considered contributions control convey corresponding count counterintuitive countweighted cover cross current customers data datasets decision deeper degree degrees degreeweighted delivered denominator dense derived described describing design designed details dhillon diameter digital directed directions discover discovered discuss disk display distance does domingos dongen dorogovtsev doyle dram draw during each edge edges eight electric electrical electricity else employ employed encoded engineering ercentage erent europ evaluation even evolution exact exceeded except exception expand expanded expanding expansion expansions experience experimental experiments explains exploration external fact factor faloutsos feasible finally first flake flow focus focusing following follows formulation framework from function functions gaertler gansner gcand generally generation generator gibson giles girvan give given gives good goodness graph graphs groups gruhl gure handbooks have haveliwala haven heavily help here heuristics high higher holland horribly however hresholds hypermedia hypertext idealized identi ieee important improved inappropriate include increases increasing individuals inferring information initially inprocess instead interestingly internet introduce introduced intuition intuitive irregular ject jeong journal justi karypis kempe kleinberg know known korea kumar large larger lawrence lead length lengths less level lgorithm liben library like likely limit limited link live lncs located logarithm longer lower lowresource machine make mallela management massive mathematical mathematicians maximizing measure measured measures media medium memory mendes mentioned methods meyer might minimum mining model models modha monma monotonically more most motivate motwani much multilevel multiplication multiplicative must nally natural naturally nature network networks never newman next ninth nition nitions node nodes normal north notion novel nowell number numerator observe occur ondition open operations optimization order organization other others otherwise over overall page pagerank pages pair palmer paper parallel parameter parentheses particular partitioning path paths pattanayak pending perform performs personalized physics pickheuristic poor posing possible possibly power practice prediction preference preferred presence probabilty problem proc product proportional proposed queries quite raghavan random ranking rather reach reachable reaching reason reasonable recall recommendations references regarding region regions relationship relationships relatively remainder report represent representing represents require requires research resistance resource resources result resulting results review richardson root roughly rows running runtime ruzzo scale scaling science scientists search second seconds section self sense sensitive seoul sept sequence setting settings several shorter shortest should show showed shown shows siam sigcomm signi similarity simplest simply simulation since single sink smal small smaller smolensky snell social software some somewhat soon source south sparse speci split spread springer stanford stoc stoer stop stopping stoppingc straightforward strong stronger structure such suggesting suggests survivable symp system systems table tages takeaway taking tardos technical technologies tend tends term terminates termination terms text than that then theoretic there therefore these thesis they this those three threshold thresholds through thus time times timing timings tiwari together tomkins topic topology total track traditional treats tree tschel tuning turning types typically uence under uniformly universal university unusual usage used utrecht value values various verlag vertices very virtanen visualization voltage volume wagner walk walks wallclock washington ways webfountain weight weighted weighting weights well were what when where which while wide widom will winograd with within without work world would york zien http://doi.acm.org/10.1145/1014052.1014078 24 A Graph-Theoretic Approach to Extract Storylines from Search Results algorithms allerton american analysis anatomy annual approach authoritative based bharat blalek bottleneck brin browsing calishain certain chakrabarti cluster clustering clusterings collections communication communities completeness computer computing conference control cutting cyber data deerwester development dhillon discovery distillation document documents dumais dynamic emerging engine engines environment etzioni fast features foundations fumas functions gather generalizations gibson good harshman henzinger http hyperlinked hypertext hypertextual impossibility improved indexing information interface international intuitive journal kannan karger karp kleinberg knowledge kumar landauer large latent letters link llrx ltering madani mallela matching mathematical maximum method mining modha networks page pages pedersen pereira problem proc proceedings processing raghavan rajagopalan ranking references research resource results retrieval scale scatter science search semantic sigir society some sources spectral stockmeyer symposium theoretic thesis tishby tomkins topic toronto trawling tsaparas tuckey university vazirani vempala vetta wide with workshop world zamir http://doi.acm.org/10.1145/1014052.1014108 53 V-Miner: Using Enhanced Parallel Coordinates to Mine Product Design and Test Data 1 added adjacent advantage after agrawal aircraft alexakis algorithm algorithms allows also amount analysis analyze ankerst application applications approach appropriate areas association baeza baker based been berchtold blindly both categorical classical clustering collected computer conclusions concurrently conf conference configuration consequently coordinate coordinates cycle data database deftereos dependent described design desirable develop dimensional dimensions dimsdale discovered discovery displayed distance done down drop easily easy edit electronic electronics enables engine engineers enhanced experimental exploration extensions fast feedback figures filter first focused forum found francisco from functionalities furthermore future generates geometry goel graphics grossman group guide haftka have hellerstein histories however identified ieee implements important improve included indicate information infovis inselberg instead intelligent interesting international intl into introduced july kaufmann keim knowledge kriegel lane large learning machine manufacturing mason matching methods mined miner mining mobile more morgan most motorola multi multidimensional narrow nelson ntts october ordering other packaging paper parallel parcovi patterns performed phone powerful previous problem problems proceedings process product products program proposed psaila query querying quinlan rearrange references results rules samiotakis second sections sequence series shaffer shapes show showing sigir similar similarity singh software specify srikant statistics string summarize survey symposium system technique techniques technologies test testing than that this thus time tirpak together tool tools traditionally trans trends used user using variables viewed visdb visual visualiser visualization visualizing visually vizcraft vldb vminer watson which wimmers winviz with work xiao yates year zait zhou http://doi.acm.org/10.1145/1014052.1014149 93 Redundancy Based Feature Selection for Microarray Data academic adorjan analysis arti based bioinformatics biomarker boston brown cancer cation cial classi coherent complementary conference correlation dash data davis dimensional discovery discretization eger empirical enabling exploration expression fang fast feature frank from gene genes genome genomic high hussain identi implementations informative intel interactive international irrelevant java jiang john jordan karp kaufmann kluwer know kohavi koller kononenko learning ledge ligence lter machine methylation microarray mining model monitoring morgan motoda olek optimal pages patterns phenotypes piepenbrock pracitcal problem proc proceedings programs publishers quantitative quinlan references relief relieff research robnik sahami schena science selection series shalon sigkdd sikonja solution subset tang technique techniques theoretical time tools toward with witten wrappers xing xiong zhang zhao http://doi.acm.org/10.1145/1014052.1014076 22 Web Usage Mining Based on Probabilistic Latent Semantic Analysis adapted adaptive advances aggregate algebra amsterdam analysis anderson application applications april arti australia authoritative automatic automatically based beach berendt berry boulder brants brisbane browsing business canada chains chang chen cial clustering cohn collections colorado communications computer computing conference connectivity content cooley data deshpande discover discovery document documents domain domingos dumais eating editors edmonton eighth eleventh ersonalization erta ertext etzioni evaluation explorations find framework from fuzziness fuzzy heuristics hofmann hong http ictai identifying ieee improve information informs intel intelligent interaction international internet issue joshi journal july june kamdar know kong krishnan krishnapuram language latent learning lecture ledge leen level ligence linear ling link lncs logs longest lrec machine madison management markov mining missing mobasher model modeling models nakagawa nasraoui national navigation neural newp niedzwiedzinski notes novemb obrien octob ontologies organization oulos oulou pages paliouras palmas papatheodorou paper path pattern patterns perkowitz pierrakos pirolli pitkow predict prediction preparation press probabilistic probabilistically proceedings processing reconstruction references relational requests research resources retrieval review robust sarukkai schechter science segmentation segovia semantic septemb session seventeenth siam sigkdd similar site sites smith soft spain special spiliop springer spyrop srikant srivastava stanford stolle studies subsequences survey symposium synthesizing systems szczepaniak technologies their third todd tool tools topic track tresp tsochantaridis usage usenix user using verlag washington website weld wide with world yang http://doi.acm.org/10.1145/1014052.1014127 71 Improved Robustness of Signature-Based Near-Replica Detection via Lexicon Randomization dovs dphulfdq edloh eulq eurzq exfnoh fduglh fkdoohqjh fkrzgkxu flhqw flnp foxvwhulqj foxvwhuv froohfwlrq frpphufh frps frpsxwlqj frqihuhqfh frqjuhvv frqudg frqwdlqphqw frrshu fudvzhoo fxphqw fxphqwv fxpplqjv gdwd gdwdedvhv gdwdvhwv gdylv ghsduwphqw ghwhfw ghwhfwlqj ghwhfwlrq gljlwdo glphqvlrqv grfxphqw grfxphqwv grhv gxsolfdwh gxsolfdwhv gxsolfdwlrq hfwru hohfwurqlf hohyhqwk hpsluh hqglxp hqjlqhhulqj hqylurqphqw hulphqwv hyroxwlrq idvw idzfhww ihwwhuo ilqglqj iulhghu iurp jdufld jlrqlv jodvjrz jodvvpdq judkdp jurvvpdq kdvklqj kdyholzdod kdzdll kdznlqj khlqw kljk krdg lfpo lghqwli lpedodqfh lpedodqfhg lqgh lqirupdwlrq lqvn lqwhuqdwlrqdo lqyhuwhg mrun mrxuqdo ndxipdqq odujh odwlq ohduqlqj olnh oqfv owhulqj pdqdjhphqw pdqdvvh pduglvd pffde phfkdqlvpv phonry phwkr phwkrg plqlqj plwud prolqd prujdq prwzdql pxowl qdplf qhdu qjhusulqwlqj qlvw qryho qryhpe qwdfwlf references rqolqh ryhuylhz sdjhv skdvh slhufh sodjldulvhg sorudwlrqv sureohp surfhhglqj surfhhglqjv surfhvv sxus uholdelolw uhsolfdv uhvhpeodqfh uhwulhydo uhxwhuv vdqghuvrq vfdodeoh vfkule vflhqfh vflhqfhv vhdufk vhjdorylfk vhtv vhtxhqfhv vklydnxpdu vljlu vljprg vljqdwxuh vlplodu vlplodulw vpduw vrerur vrflhw vsdp vsdpphuv vwdwlvwlfv vwhp vwhpv whfkqlfdo whfkqltxhv whfkqrorj whvw wlsvwhu wudfn wudqvdfwlrqv wuhf xqlyhuvlw xvhql yhuvlrqhg ylyr yoge zdjvwd zheb zhege zhlj zlgh zlwk zrunvkrs zruog http://doi.acm.org/10.1145/1014052.1014099 44 A Rank Sum Test Method for Informative Gene Discovery ∗ academy accuracy activities advances alon american analyse analysis annual arrays association asymptotic august availabe barkai barnhill based bednarski bioinformatics biology biometrics breast broad cambrige cance cancer cation cdna chris christianini ciency class classi clustering clusters coherent colon combinations compact comparison comparisons computational computer conference criterion data derisi dimensional ding discovering discovery discriminant discrimination discriminative distributions donna drug duda dudoit edition editor eric expression extreme feature features figure fourth francisco fridyand from functional furey gene genes genetics genome genomewww golub goulden government grosse grundy guyon hart hauessler heterogeneous highly holden html http human hwanjo ieee index individual industry informative international isabelle jason ject jiang jiawei jill jinsong jiong joachims john journal kernel know lander large leaf learning ledge lehmann leukemia level levels machine machines making mesirov metho method methods microarray mining molecular monitoring multiple national nature nikitin noble normal number oligonucleotide oncology ordering pablo pages paper parametric parametrics pattern patterns paul pavlidis penland practical prediction press princeton probed proceedings ramanathan rank ranking ranks references revealed samples sbcmp scale scholkoph schummer science sciences seattle selected selection seventh shtml sigkdd signi significance sixth slonim society sons speed stanford statistic statistical stephen stork support table tamayo tang test tests three tissue tissues todd track tumor university using validation value vapnik vector versus vladimir wang weston wilcoxon wiley william yang ydata york zhang http://doi.acm.org/10.1145/1014052.1014063 9 Data Mining in Metric Space: An Empirical Analysis of Supervised Learning Performance Criteria aaai above absolute accuracy acknowledgments actually advances advertisement advertising after agent airborne alex algorithm algorithms also analysis anything appendix applicable applied approximated approximates area arti assumptions ativ average averaging aviris backprop bartlett based baseline bayes bayesian because become better between binary binned bins blake boost boosted boosting break buntine calculate calculated calibration cart cases cation characteristic chettri cial city class classi client clients collaborators combine community compared comparison compatible compute computed conference consider continuous convert correct cost creators crew criteria criterion cromp cross cult curve data databases dataset days decay decision degree degroot descent described description deviate diagrams discovery distance distinguish distributions domingos each early editors eighth elds eleven entropy epochs equal erence erent error errors euclidean evaluation even evenly everything examine example examples experiments factors false fawcett feng fienberg flach following follows forecasters fraction frequency from gain generate geometry geoscience giudici given gives gradient gualtieri harmonic have help hidden hypothesis icml imprecise induction information instead intel interactive interested international isometrics january joachims john johnson june kernel kernels king know ksikes labels laplacian large learning least ledge level levels lift ligence likelihood likely linear locally machine machines make makes making many margin marketing maximized maximum mean measures medicine merz method methods metric metrics minimizing minimum mining mmlls models momentum more most much name negatives nets neural next nice nition number observed often only operating optimal options order other output outputs over package pages paper parameter percentage performance platt plot plots point points polynomial poor popular population positiv positive positives possible potential practical precision precisions predict predicted predicting prediction predictions predicts press principled probabilistic probabilities probability probably problems proc properties proportion proven provost pruning radar radial randomly range ranging rankings rate ratio real recall receiver recently references regression regularization regularized relative reliability repository research respectively respond retrieval rmse root roots running same scale scaling schoelkopf schuurmans score send sensitivity sent sets setting settings seven shutherland sigmoid similar single size smml smola smoothing software some sons space spaced speci splitting squared stanford statistic statistician statlog steps stmp stopping stumps such suggests summary support suppose svmlight svms take target targets than thanks that their them then there these this threshold thresholds through times together track train trained trainset tree trees trivially true tting type under understanding units used usually validation value values variations vary vector visualization wants ways weight weighted when where widely width widths wiley will with without workshop world would wwii xgobi xgvis york http://doi.acm.org/10.1145/1014052.1014066 12 Adversarial Classification aaai adaptive advanced aerial agent analysis androutsopoulos anti application approach arms arti attacks based bayes bayesian been bibliographies bibliography boston brady brunswick cambridge canada case categorization cation chan cial classi cogsci comparison conference cost data database datasets detecting detection discovered discovery dumais edmonton eighth eleventh email engine english environments event evolution fawcett feature florida framework games gazumped google guide heckerman high horvitz html http ieee image impact imprecise information institute intel international john june junk karkaletsis kaufmann kluwer know knowledge kohavi krebs language large learning ledge lexical ligence ligent lists littman lloyd ltering machine madison mahoney mail mailing make management markov mccallum members memory mining mkearns models morgan multi naive nasd network nielsen nigam nips nongnu nonstationary normal novel november ongoing online organization ottawa pages paliouras pennsylvania peter philadelphia piracy post press princeton proceedings provost race regulation regulatory reinforcement rennie retrieval robertson robust rogers sahami sakkis search selection senator sense sensitive sigkdd sixth smith spam spurs spyropoulos stamatopoulos study subset surveillance system systems tech technology text theory trying turney tutorial university update upenn volume washington wordnet workshop wrappers http://doi.acm.org/10.1145/1014052.1014065 11 Exploiting Dictionaries in Named Entity Extraction: Combining Semi-Markov Extraction Processes and Data Integration Methods aaai abstracts account acids acknowledgments active addition adds adjacent advanced agency agichtein alberta algorithm algorithms allowing allows also alto altun always analysis answer applications approximate approximating architectures arguably army arti articles association assumptions attributes august austin authors automatic automating autonomous available awkward barabara based baseline bases behavior bhamidipaty bikel biocomput biocomputing biological biology biomedical blast bollacker boot bootstrapp borkar borthwick bottom brunswick bunescu business cali california cambridge canada candidate carnegie cases cation change cial citation classes classi classifying cohen collections college collins combine comments compared comparison computational computer conditional conf conference conll constructing construction continuum contract coordination corpora correct corresp coupling crammer craven cult darpa data database dataprivacy datasets decemb decoding deduplication defense demetriou dence deshmukh design dictionaries dictionary digital direct discovery discriminative distance distributional diverse dramatically durban dynamic ecause ective eddy edmonton eginning eighth elds eling elsewhere emnlp empirical encoding endent entire entities entity entitylevel entries entropy enzyme eople erate erceptron erences erformance erforms eriments erty ervised espinosa estimation etween examples existing exploiting extending extent external extract extracted extracting extraction features fienb fifteenth finding fluck formalism formalize formulation foundation freitag freund friedman from funded fussell gaizauskas game gene giles goal grant grants gravano greatly grishman hanisch hidden high hofmann however html http humphreys icml identifying ieee iiweb ijcai improved incorp indep indexing information instance institute integrated integration intel interactions interactive international internet ipto irrelevant irvine ismb isri java ject jects jersey joint jones journal july kate kaufmann klein know knowledge kraut krauthammer krogh kumlien labeled language large lawrence learing learn learning learns ledge length lerch level libraries ligence ligent linear linguistics linkage lists littlestone mach machine machines make malouf management manning many marcotte margin markov matched matching maximum mccallum measures medline mellon merely methods metrics mevissen mining mismatch mitchison model models molecular mooney more morgan morozov multi multiclass multistrategy must naacl name named names national natural nder negative nigam note nucleic numb onding online only orating organizational ortals orted osed ositive other ound outp output paci page pages palo paper papers park parse parsing part participate particular pattern pereira plaintext planning playing preparation press previous privacy probabilistic problem problems proc proceedings process processing programming prop protein proteins provides publication queries quickly ramani random rather ratnaparkhi ravikumar reacting recognition record records references relational relations relax rennie research retrieval reviewers rilo rosenfeld roster rules rzhetsky santa sarawagi schapire scheme school schwartz science sciences scienti score secondstring segmental segmentation segmenting segments semantic semi sequence sequential sequentially seventh several seymore shallow sigdat sigkdd sigmod similarity simulated singer single sixteenth sixth snowball something sometimes source sourceforge sources state statistical sterling strapping string strong structure structured structures substantive successful such supp surprisingly survey sutton sweeney symp symposium systems tags tasks teams technical techniques technology texas text than thank that their theory these thesis they this threshold toolkit track training tsochantaridis ultimate ultraconservative university unlabeled unpublished unstructured unsup useful users uses using usual utexas vector version versus very washington weischedel well what when whether which while wiley williams winkler winston with wong word words workshop zimmer http://doi.acm.org/10.1145/1014052.1014105 50 Learning to Detect Malicious Executables in the Wild aaai academic advanced advances aegis after against aiken alamitos albert algorithm algorithms alternative analysis annals anonymous anticipate application approach approaches arcing area arises arnold arti association attacking auditing august author authors authorship backdoor bacon bagging bartlett based bauer bayesian berkeley berlin best biannual bioinformatics biologically birmingham boosted boosting boser boston breiman burges california called cambridge categorization cation characteristics chess chicago christodorescu cial class classi classifying code cohen combined company comparative comparison compression computational computer computers computing concluding conference considered constructing corporation cost council curve data decision defenses delivered delivery department detect detecting detection determine diagnostic dietterich discovery discrimination document does domingos drummond duin dumais durning each ective edition editors eighteenth elded empirical ensemble ensembles entail environments eskin european evaluating evaluation example executable executables expected experimental experiments expert experts explicitly extending fast fawcett feature features forensic forensics fourteenth fourth francisco frank freund frieder from functional functions furthermore future gaithersburg gove gray grossman guyon hand have heckerman heuristics hexdump hill holte hope horvitz html http identifying ieee image implementations important imprecise improve indeed indexing indianapolis induction inductive industry information infosec inspired instance institute intel international investigate jain jankowitz java jiang joachims john joint journal junk kaufmann kephart kernel kibler kjell kluwer know known kohavi krintz krsul kuhns kurt labels laboratories lafayette large lawrence learning ledge levitt ligence linguists loss lter ltering macdonell machine machines maclin macmahon mail malicious malware management mannila many margin maron mass master maximum mcbride mcgraw mecs menlo methodology methods metz mika miller minimal mining mitchell mitre modify morgan morisett moss most multiple must national networks neural nishikawa obfuscation october olsson open opitz optimal optimality optimization overall page pages paper papers park pascal pattern patterns pazzani pederson perform pickett plagiarism plan platt popular possible practical press principles privacy probabilistic probabilities problem proceedings processing produced program programs provided provost publishers publishing purdue pursuing quinlan radiologic randomization recognition references relevance relevant remarks removing report representation representations representing research resulted results retrieval review robust rossmann rule sahami sallis sams schapire scheme scholkopf schultz schuurmans science search security selection september sequential seventh shake sigkdd signal signatures simple sixth smola smyth softeng software soman sorkin speare springer standards static statistical statistics step stolfo strategy student study such suggest support swets symposium system systems task technical techniques technology tenth tesauro text that theory thesis third thirteenth this three tool tools track trackrnment training transactions trees tuwien twelfth under university unknown usenix using valuable vapnik variants variety vector verlag vigna virtual virus viruses visualization voting weeber west when which white wild will with witten woods work workshop yang york zadok zero http://doi.acm.org/10.1145/1014052.1016925 111 Analytical View of Business Data accomplish aggregated aggregation alejandra alive allowing analysis analytical application applications april array based bianalysis botto break bring brookins business classes colorado complete component conclusion consistency consuming context customers data decisions default demand detail dimensional during dwsqlsy edition employees enable engine enterprise entities evaluation exploit extending extension files following framew framework from fully functionality garrido guide gustavo have http hypermedia implemented infers information intellidrill intelligence john kena kimball logical main make metadata microsoft mission model modeling models multi navigate needed next notion object olap oriented page paper participate point points potential presented preserved process prodtechnol programming providing ralph realize references related report reporting resource result rossi scenarios second server service single sitestream sons subsystems team technet their this three through toolkit transformation treeview unleashed using utilizes warehouse warehousing which wiley will with yukon http://doi.acm.org/10.1145/1014052.1014093 39 IDR/QR: An Incremental Dimension Reduction Algorithm via Q R D ecomposition abbadi academic accurate agrawal algorithm algorithms american analysis applications applied arti association baltimore based batch belhumeour benavente berchtold black cation centroids chakrabarti chandrasekaran chatterjee cial class classi clustered comparison component computation computations computer computing computional cost covariance criterion cviu cybernetics daniel data database databases decomposition dimension dimensional dimensionality discriminant duda dynamic edition eigenfaces eigenfeatures eigenspace etween extraction face factorization fast feature features figure fisherfaces friedman fukunaga generalized gmip golub gragg gram graphical hall hart hespanha high honavar hong hopkins howland iccv ieee image improving incremental index intel introduction jain janardan jection jections jeon johns jolli jonathan journal kanth kaufman keim kong kriegman krzanowski learn learning least ligence linear loan lower machine manjunath marshall martin martinez mathematics matrices matrix mccarthy merging methods models multimedia multiple multivariate networks neural optimization organizing pages paper park pattern percentage performance polikar preserving press principal problems processing ravi recognition reduction references regularized reorthogonalization report representation research retrieval robust rosen roychowdhury schmidt searching seconds self separability siam similarity singh singular soundalgekar spaces speci spectroscopic splitting springer squares stable statistical statistics stewart stork structure structures supervised surveys swets systems technical text third thomas time torre track training trans udpa undersampled understanding university update updating using value verlag versus vision vldb volume wang weng wiley winkeler with york zhang http://doi.acm.org/10.1145/1014052.1014113 57 Column-Generation Boosting Methods for Mixture of Kernels aaai advances algorithm alignment arenas bartlett basis becker bennett bishop boosting bousquet burges cambridge classi column comparing complexity conference controlled crammer cristianini data demiriz design discovery editors elements elisseef embrechts ensembles error fawcett figueiras friedman function garca garcia gaussian generalized generation ghaoui girosi growing hastie hernandez herrmann heterogeneous hill hyperkernels hypothesis ieee inference information international jective jimenez john jordan journal kernel kernels keshet know lanckriet large learning ledge linear lkopf machine machines mangasarian margin mark matrix mcgraw menlo mining mishra models momma mora multi nash navia networks neural neurocolt neurocomputing ning nite niyogi nonlinear obermayer optimizing oriented oxford pages park parrado pattern percentage poggio prediction press problem proceedings processing programming radial recognition references regression report research schuurmans semide shawe sigkdd signal singer smola sofer sons spaces sparse springer statistical sung support svms systems taylor technical theory tibshirani transactions tsch twentieth university using vapnik vazquez vector vidal wiley williamson with york http://doi.acm.org/10.1145/1014052.1014122 66 Discovering Additive Structure in Black Box Functions accessible additional additive agrawal algorithm algorithms allows analysis annals anova approximation areas association asymptotically bagging black boosting boston bound breiman buja carlo categorical cient class classi clean complexity components comput computers conclusions conference cross cuto data databases dean decomposition demand demonstrated dependence developed diagnostics dimension dimensional dimensionality ding directly display distribution distributions economics empirical employs environmental estimates estimating estimation estimators evaluations exploration extending extrapolation fast feinberg figure friedman function functional functions global good gradient graphical greatest greedy harrison have hedonic here hierarchical high hofmann hooker housing html http index indication indices interact interaction interactive international interpretational intrinsic ject jiang journal large lattice learn learned littman lower mach machine make management manner math mathematical mathematics mean measure mining models monotonicity monte more multidimensional network neural nonadditive nonlinear normal only owen partial plots predictor predictors presented press prices problem proceedings properties provide providing quadrature quasi references regression report represent research results roosen rubinfeld rules sampling scaling search sensitivity shrinkage simul simulation sinica size sobol srikant stanford stat statistica statistics swayne system technical terms test that their theory these thesis this through tied trained university used user using variable very viewed visualization which with work xgobi xgvis http://doi.acm.org/10.1145/1014052.1014060 6 Fast Nonlinear Regression via Eigenimages Applied to Galactic Morphology accepted actual additive adelman advances agreement algorithm also anderson andrew angle apparent appear appendix area argument arti assume astron astronomical astrophysical away axis background ball because been before being benjamin between blurred both bottleneck brief brightness brinkmann brunner bulge cadez cation center christopher cial clarity class classi classic classical classify clockwise combination commonly compact comparison component components consists constant continuous created database david decomposition deep degree delta describe describes description detailed detector digital digitized dimensional discretization discussion disk disks distributions each effect egrees elements ellipticity elsewhere empirical entire equation error expected experiments expert exponential exponentially faber figure fitted form formulae found freeman from fukugita function functions galactic galaxies galaxy garth gbul gmorph grif groth have henceforth here horizontal hubble human illingworth image images impey inclination indicates information integrated integrating into irregular journal kamath kaufmann kirshner known kormendy lahav learning lengths linear mathworks matlab matrix medium method model modeled monthly morgan morphology most myungshin nakamura nebulae networks neural nicole nitely noise normal notices number objective observed offset okamura omit omitted only onto ostrander over paper parameter parameters peng perhaps phillips photons pixel pixels plane presentation press processing projected pure quartiles radius ratio ratnatunga realm recorded recovered refer references relative release research returned rotation royal sarajedini scale sersic shape shapes simard simpli sloan slonim smyth society somerville space spectra spheroid spiral strip structural summary supervised supplements surface survey systems takes tapering task technical telescope that their they thin this those three tishby total toward towards track tted types unblurred unit units university used using value variables varies vaucouleurs vector version vertical vicki viewer vogt volume weiner wfpc when where which widely will willmer with yale york http://doi.acm.org/10.1145/1014052.1016919 105 Tracking Dynamics of Topic Trends Using a Finite Mixture Model algorithm algorithms allen analysis anomalous applications approach automatic based bursty cambridge cant capabilities carb cation changes classi cluster clustering coding complexity conditioned conference decision detecting detection discounting document ective editor ehaviors ervised estimation event extension extracting features finite forum from generation gong graphical harve hetzler hierarchical hinton ieee incremental inform information international japanese john jordan justi kleib know lavrenko learning ledge letters line lists management matsunaga mclahlan milne mixture mixtures model models neal nement nite norwell novelty onell other outlier over overview papka peel pierce press probability proc proceedings proceesings processing references retrieval retrosp rissanen selection series sigir signi sons sparse statistics stochastic streams structure study swan symposium takeuchi technologies technology text that theme themeriver theoretic theory time timelines topic tracking trans universal unsup using variants varying view visualization visualizing wiley williams with yamanishi yang zang http://doi.acm.org/10.1145/1014052.1014107 52 Density-Based Spam Detector adachi addison analyzes annual anti application architecture arti based bayesian better billion bytes caching characteristics checksum cial classi clearinghouse cluster compression computational computer computing conclusion conference contents continuous density desk detect detection development dion distributed distributions document edition editor eken eleventh enough erformance estimating fast filtering generic good goverdustny graham hall hanafusa handbook handle hashing high hill hosakowa html http ietf industry information innme intel introduction invariance ipsj izumi japanese java john kaufmann known knuth langley large learning ligence lindb linguistics lter ltering machine machines mail mails march mass mateo mcgill mcgraw memory method methods minimal mitchell mitkov modern morgan mori motoda mtas much nishikawa optimization organization orleans orts osed over oxford pages paper pentium platt predictive prentice press proc proceedings processing programming programs prop proxy publishers quinlan radio raita recommendations references report require resources retrieval rhyolite robinson saito salton search searching second sequential server single small smtp society software sorting space spam spamassassin spamdetection stories structured supp support tanenbaum technical tend teuhola text this though track training transformation tsuji uehara uncertainty useast using vector wada waikato washio weblogs weka wesley which wide with world yoshida http://doi.acm.org/10.1145/1014052.1014067 13 Regularized Multi–Task Learning academic adaptive advances aggregate algorithm allenby allow annals application applied approach arora arti bakker based baxter bayes bayesian berkeley bias boussios breiman brown california canada cardinality caruana categorization choice cial classics clustering collection colt combining common computation computational computer conjoint connection consider consumer could cristianini cross customization dahan data datamining david degree demand dept discovery disparate division dynamic each ecomes econometrics editors empirical equation erent eriments erts estimates estimation even evgeniou example exploiting fast features framework friedman from function functions fusion gating gehrke generalized genomic ginter hauser have heisele heskes heterogeneity hierarchical higher icml immediate improving index individual inductive information insead intel jacobs ject jordan journal july kernel kernels kluwer knowledge lanckriet langley learn learning ligence linear machine machines management mangasarian marketing mathematics matrix means measure membrane mercer micchelli mixtures model modeling models multi multiple multivariate networks neural noble nonlinear note novemb observational olynomial onses ossibility paper parallel parameter parts philadelphia poggio polyhedral pontil pool pratt predicting prediction primary problem proceedings processing programming protein publishers rates references regression regularization relatedness research resp ridge robust rossi royal sampling school schuller science secondary selective septemb series serre sets siam silver simester sloan society sources splines statistical statistics sullivan summer supp systems target task tasks technical that theoretic theoretical theory these thrun toubia track transfer university used uses using valued vancouver vapnik variety vector vetter wahba ways where wiley wish working would york zacharia zidek http://doi.acm.org/10.1145/1014052.1014139 83 Privacy Preserving Regression Modelling Via Distributed Computation analysis applications applied april association atallah august available based berlin berthold bibliography borgelt bradley calculating cambridge cation chen cient classi clifton cloudcroft clustering collective computation computational computer computing conference data databases decemb derivatives dgii discovery distributed duggan edition editors eighth erger explorations flannery function german graphical hershb heterogeneous html intel international iternational journal july kargupta karr know kruse ledge lenz ligent line linear lncs lyons madden means method minimum mining multi multivariate nding ninth niss numerical orts over pages paradigms paral partitioned party powell preserving press privacy problems proceedings publication recipes regression reiter review rule sanil scienti second secure security septemb sets several siam sigkdd springer statistical statistics submitted symposium technicalrep teulosky their tools university using vaidya variables vertically vetterling wavelet weisb wiley without workshop xing http://doi.acm.org/10.1145/1014052.1016912 98 Cross Channel Optimized Marketing by Reinforcement Learning advanced advances advantage after again approach approximate approximately approximation apte arlington arti assumption attributable baird barto based batch bradtke california campb cantly change channel cial conclusions conference continuous cost course cross current data dating decision declining diego direct discovery distribution diverging does eaks ecomes ected ective ehavior eighth elnieks empirical employing encouraging ercent estimators evaluation even existing figure from function graphs imply increase information initialized intel international introduction investi iterations journal july june kaelbling kakade know langford lead learing learning ledge less life ligence limitation littman machine mailing making march marketing markov massive measure measures mentioned method methodology methods mining modeling moore more much natara nature nelson networks neural nite noting observed obtained olicies olicy optimal optimized output over pages part pednault policies policy press problem problems proceedings processing progresses quite read real references regression reinforcement relatively reliable research resp results saks sampling schroko second segmentation segmented sensitive sequential sets seventh show shows siam sigkdd signi similar since start starts state store survey sutton systems targeted technical that then therefore these thesis third this thought time tipu trend typical uninformative university updating used using validated value verma virginia volume wang were when which with world worth zadrozny http://doi.acm.org/10.1145/1014052.1014062 8 A Probabilistic Framework for Semi-Supervised Clustering aaai academy active adaptive advances algorithm algorithms analysis annie annual application approaches approximation approximations arti assignment association background baeza banerjee bansal based basu bayesian bennett berkeley besag best bilenko bioinformatics bliek blum botstein boykov brady brain bregman brown cambridge cardie caruana cation cell center chawla cial cient classi clifford cluster clustering cohn combining comparing computational computer concept conference constrained constraints continuum cornell correlation cover cvpr data decompositions demiriz dempster detection dhillon directional dirty discovering discovery display distance distributional distributions divergences documents duplicate edition editor eisen eisenberg elds elements embrechts engineering english equivalence expression external feedback focs foundations from functions geman gene generative genetic genome ghosh gibbs graphical graphs guan hammersley hard hertz heuristic hidden hillel hinton hochbaum icdm icml ieee ijcai images imaging impact incomplete incremental inference information instance intelligence intelligent interaction international interscience joachims john joint jordan journal july jupp justi kamvar kaufmann kearns klein kleinberg knowledge koller labeled labeling laird large lattices learnable learning level likelihood linguistics localizing machine machines macqueen making manning mansour manuscript marcotte mardia markov mateo mathematical mathematics maximum mccallum means measure measures medical meeting merugu methodological methods metric mining mitchell model models modern modha molecular mooney morgan most multivariate national neal neto networks neural nigam nineteenth ninth nite observations occurrence operations other page pages pairwise pathways pattern patterns pearl pereira phylogenetic pictures plausible possible press prior probabilistic probability problem problems proceedings processing protein proteins random reasoning recognition references relations relationships relaxation report research restoration retrieval ribeiro rogers royal rubin russell schroedl science sciences search seeding segal segmentation semi series seventeenth shental shmoys siam side sigkdd similarity sixteenth smith society soft some sons space sparse spectral spellman statistical statistics stochastic strehl string supervised supervision support symposium systems tardos technical text that their theoretic theory third thomas thrun tishby training transactions transductive uncertainty unifying university unlabeled unpublished user using validity variants vector veksler view vision wagstaff wang weinshall wide wiley with words workshop xenarios xing yates york zabih zhang http://doi.acm.org/10.1145/1014052.1014106 51 Predicting Prostate Cancer Recurrence via Maximizing the Concordance Index accuracy advancing after analysis approach arti attikiouzel australian biganzoli biology boracchi branford breast brown burke cagiannos cancer censored choong cial clin comput data davidov desilva detection discov discovery drug epidemiology erative feed forward goodman graefen green gronb holland ieee improve john journal karakiewicz lancet localized logistic mariani marple moran naylor networks neural nomogram oncol ostop partial prediction prognosis prostate prostatectomy psychophysics radical recurrence references regression rosen signal sons stat study survival swets systems theory through today trans validation wiley york http://doi.acm.org/10.1145/1014052.1014118 62 Kernel k-means, Spectral Clustering and Normalized Cuts academic algorithm algorithms allows also although analysis annual anti applications approach area areas augmented august authors bach banerjee based bioinformatics bregman cambridge case cient clustering clusterings clusters collections component computation computationally computations computer conference considers correlated cristianini cuts data describ dhillon diametrical dimensional ding discrete discusses divergence document eared ecially ectral eigenvalue eigenvector encompass engineering enhance erent factorization feature focusing foundations from function gene generalize ghosh girolami golub good gram guan have high hint hopkins identifying ieee image info inspired intel international introduction issue iterative jective johns jordan kannan kernel kernels kluwer kogan large largely learning ligence linear lkopf loan local machine machines malik marcotte matrix maxi maximization means mercer merugu method methods mining muller multiclass nding networks neural nips nite nonlinear normalized notion numb objective original other pages pattern plus presented press problem proc proceeding proceedings processing publishers recast recent references related relation relaxation results roshan science scienti search segmentation semi september shawe show siam simon smola space spectral such supp support symposium systems takes taylor text than that their they thus time trace traditional trans transactions treat university used using vector vempala version very vetta vision weighted weiss well when which with work worse http://doi.acm.org/10.1145/1014052.1014146 90 Mining Scale-free Networks using Geodesic Clustering academic algorithmic algorithms analysis analyzing anatomy applications applied approximates approximation authoritative barab based been believe bianconi birch bollob brin called cally cambridge cation cells centrality characteristics cient classi classifying clustering clusters community complex computations computing conclusion conf conference connections contracting could crpc customers data databases deling described designed detail detecting developed discovered discovery distribution djoko domingos done drezner dual edition egel eighth elsevier engine environment erformance erlinked estimating eurographics example exploit explor facility faust figure fisher form found framework free from function future garland geodesics graph graphs gspan hamacher handbook have high holder however hypertextual icdm idea important interaction international intl introduced irvine java jeong june jung karypis kaufmann kleinb kluwer know knowledge krebs kumar large ledge lethality levels livny location madadhain management many mapping marketing mason mathematical mesh metho mine mining mirkin modeling morgan moto much multilevel multiple multiresolution nature network networks newman newsl ninth noted oltvai only ortance ortunities other page pages paral partitioning past pattern patterns performing perspective phys power press problem problems proc proceedings properties protein publishers ramakrishnan random real reduction references relative report representative review richardson scale schlo science scienti search second sept seventh sharing showed siam sigkdd sigmod simplify simulations since sites smyth social sources sparse speci speeding springer state structure structures substructure such survey system technical techniques terrorist that theory these this thus toward underlying understanding universal university useful using value vertices very viral visual visualizing volume want washio wasserman when white wide work workshop world years zhang http://doi.acm.org/10.1145/1014052.1014075 21 Mining the Space of Graph Properties agrawal algorithms arvind asso association banks bhalotia brin browsing california chakrabarti charuta chile ciation citeseer clubnexus conference data database databases dbpubs dehasp discovery edbt engineering extending extracting fast february from gaurav germany hannu heidelb html http hulgeri icde international jose keyword large march mining pages patterns proceedings rakesh ramakrishnan references relational relations rules santiago searching septemb sergey soumen spain springer srikant stanford sudarshan technology toivonen using valencia verlag very vldb webdb wide workshop world http://doi.acm.org/10.1145/1014052.1014061 7 Clustering Time Series from ARMA Models with Clipped Data acute akaike algorithm allow analysis analyzing anglia annual antennal approach arma arti atlantic attractor audio australasian autoregressive average bagnall based batistakis benelux biology biometrika bishop blender brain broerson burg cadez cation chapter cial city classi clipped clustering computation computational computing conference continuous criteria curve cyclone data density dermatas direct discovery discretised driven dynamics east econometrics ects editors empirical entropy error estimation estimator exploration expression finding forecasting fraedrich frey from fuzzy galan galizia gaussian gene geophysicists gerber geva godolphin gunopulos halkidi heckerman herz identi ieee iglesia information intel international jaakkola janacek joseph journal kerem know kokkinakis large ledge ligence likelihood lobe lunkeit mannila maximum medicine meek meeting meteor mining mixture mixtures model models molecular moving navigation neural neuro ninth north odor oklahoma olfactory pages pathology pattern patterns polynomial presented press principles proc proceedings process processing quart random rapid recognition recomb references regimes regression report representation research royal sachse sample school sciences second series signal similar simon simple site sixth smyth society spectral speech state statistics symposium systems technical temporal time track university unsupervised using vazirgiannis visualization waele white with workshop zhang http://doi.acm.org/10.1145/1014052.1014151 95 A Data Mining Approach to Modeling Relationships Among Categories in Image Collection accuracy addressing advanced advantage algerbraic also american analysis annotation another archiving august automatic average based beigi benitez browsing cantly castleman categories category cation cbir chang ciency classi cluster collection comm compared comparison complexity compuer computation computer concept conclusion conf conference content correlation corresp cument cvpr data databases deerwester derived digital discovery distributed dumais dunham ectiveness engineering english entropy erformed eriment error errors evaluate evaluation every evident explicitly feature features from furnas fuzzy geometric graph grimson guration gure hall harshman have hence hidden hierarchical high honkela hoppner huang ieee image images improve indexing information intel internation international introductory invariants jects john journal june kaski klawonn know kohonen kruse kumar lagus landauer large latent ledge less ligence linear linearly lipson lling machine manjunath massive measurement metho method methods mining model moment more most motivations much multimedia networks neural novel numb observed onding online organization ositioning ositories other overall overlap paatero pages pattern precision precisions prediction prentice printed proceedings processing qbic query rate ratio recognition recogntion recorded reduced references region relevant resource results retrieval retrieved returning rigid river runkler saarela saddle salo scene scenic scheme science search second selection self semantic semantics septemb sets shannon sheikholeslami shown shows signi similarity sinha sixth smith sociation sons spie state statistics storage structural substantially substantiate system systems table taubin tech testing texture than that these they this thus time topics tractable training trans transform tree uence uncertainty using validated value video vision visual volume wang washington water wavelet where wiley with without wolf words york zabih zhang zhou