rant framework rank edit share web search case structur approach manag unstructur data challeng manag unstructur data repres largest data manag opportun commun manag relat data risk let opportun cede plai field player rang commun ai kdd ir web semant web industri player googl yahoo microsoft essai explor improv situat draw lesson learn manag relat data outlin structur approach manag unstructur data conclud discuss potenti implic approach manag kind relat data identifi field garden problem web inform monitor olap multicriteria mainten schedul optim complex extract program evolv text data inform extract approach consid static text corpora appli real world text corpora dynam evolv time extract inform date appli repeatedli consecut corpu snapshot appli scratch snapshot lot time avoid recent develop cyclex system recycl previou result speed subsequ corpu snapshot cyclex demonstr promis recycl idea work limit consid program singl blackbox practic program complex multipl blackbox connect composit workflow paper present delex system remov limit identifi difficult challeng rais delex includ model complex program recycl purpos implement recycl process effici search optim execut plan vast plan space recycl altern describ solut challeng final describ extens experi rule base learn base program real world data set demonstr util approach kiwi scalabl subspac cluster algorithm gene express analysi case structur approach manag unstructur data challeng manag unstructur data repres largest data manag opportun commun manag relat data risk let opportun cede plai field player rang commun ai kdd ir web semant web industri player googl yahoo microsoft essai explor improv situat draw lesson learn manag relat data outlin structur approach manag unstructur data conclud discuss potenti implic approach manag kind relat data identifi field build commun wikipedia machin dhuman partnership approach abstract rapid growth web commun motiv solut build commun data portal solut follow roughli approach approach libra cites cimpl emploi semi automat method extract integr data multitud data sourc approach wikipedia intellipedia deploi initi portal wiki format invit commun member revis add materi paper combin approach build commun portal hybrid machin human approach bring signific benefit achiev broader deeper coverag provid incent user contribut portal date user effort sens enabl build commun wikipedia back underli structur databas continu updat automat techniqu outlin idea approach describ challeng opportun provid initi solut final describ real world implement preliminari experi demonstr util approach inform extract challeng manag unstructur data past year build end end system wisconsin manag unstructur data extract integr user interact paper describ kei inform extract challeng run sketch solut discuss develop declar languag optim languag gener proven incorpor user feedback process develop wikibas user interfac feedback effort push rdbmss work suggest manag unstructur data open interest research challeng challeng greatli benefit wealth work manag structur data carri databas commun joint cluster analysi attribut data relationship data connect dcenter problem minimum consist subset cover problem applic data mine paper introduc studi minimum consist subset cover mcsc problem finit ground set constraint find minimum number consist subset cover subset consist satisfi mcsc problem gener tradit set cover problem minimum cliqu partit dual problem graph color instanc practic data mine problem area rule learn cluster frequent pattern mine formul mcsc instanc discuss minimum rule set problem minim model complex decis rule convers cluster problem minim number cluster satisfi distanc constraint show mcsc problem find applic frequent pattern summar mcsc formul propos graphbas gener algorithm cag directli applic cag start construct maxim optim partial solut perform driven specif gener search dynam maintain bipartit assign graph simultan learn set consist subset small cardin cover ground set experi benchmark dataset show cag achiev good result compar exist popular heurist user dcentric research challeng commun inform manag system cimpl joint project wisconsin yahoo research build system manag inform onlin commun paper discuss fundament role user plai system difficult user centric research challeng rais role respect contribut system access leverag social interact user infer nearest rectangl learn revisit abstract nearest rectangl nr learn train instanc gener hyperrectangl queri classifi class nearest rectangl method receiv attent introduct hybrid learner gain accuraci advantag sacrif classif time compar interpret eager learner decis tree paper seek accuraci improv nr learn control gener rectangl infer rectangl infer compact conserv good make local decis experi benchmark dataset valid effect propos approach turn cluster pattern rectangl dbase discrimin data descript ultim goal data mine extract knowledg massiv data knowledg ideal repres human comprehens pattern end user gain intuit insight data mine method produc readili understand knowledg cluster algorithm output set point cluster paper perform systemat studi cluster descript gener interpret pattern cluster introduc analyz descript format lead express power motiv defin descript problem trade off interpret accuraci present effect heurist algorithm empir evalu turn cluster pattern rectangl dbase discrimin data descript ultim goal data mine extract knowledg massiv data knowledg ideal repres human comprehens pattern end user gain intuit insight data mine method produc readili understand knowledg cluster algorithm output set point cluster paper perform systemat studi cluster descript gener interpret pattern cluster introduc analyz descript format lead express power motiv defin descript problem trade off interpret accuraci present effect heurist algorithm empir evalu cluster descript format problem algorithm ultim goal data mine extract knowledg massiv data knowledg ideal repres human comprehens pattern end user gain intuit insight data mine method produc readili understand knowledg cluster algorithm output set point cluster paper perform systemat studi cluster descript gener interpret pattern cluster introduc analyz descript format lead express power motiv defin descript problem trade off interpret accuraci present effect heurist algorithm empir evalu joint cluster analysi attribut data relationship data connect dcenter problem 