world nutshel concis rang queri proceed intern confer data engin nearest neighbor queri knn djoin larg relat databas free approxim string search spatial databas abstract work present index structur mhrtree effici answer approxim string match queri larg spatial databas mhr tree base tree augment min wise signatur linear hash techniqu min wise signatur index node concis represent union gram string tree analyz prune function signatur base set resembl queri string gram tree index node mhr tree support wide rang queri predic effici includ rang nearest neighbor queri discuss estim rang queri select accur present adapt algorithm find balanc partit spatial string inform store tree extens experi larg real data set demonstr effici effect approach probabilist string similar join edit distanc base string similar join fundament oper string databas increasingli applic data clean data integr scientif comput deal fuzzi inform string attribut intens effort devot process determinist string join manag probabilist data model process probabilist string larg unexplor territori work studi string join problem probabilist string databas expect edit distanc eed similar measur discuss probabilist string model captur fuzzi string valu real world applic string level model complet expens repres process charact level model succinct represent uncertainti string exist posit comput eed probabilist string prohibit expens design effici effect prune techniqu easili implement exist relat databas engin model extens experi real data demonstr order magnitud improv approach baselin log footstep quantil summari entir histori quantil crucial type order statist databas extens research focus maintain space effici structur approxim quantil comput underli dataset updat exist solut design support current updat snapshot dataset queri past version data answer paper studi problem histor quantil search object enabl approxim quantil retriev snapshot dataset histori problem import analyz evolut distribut monitor qualiti servic queri optim tempor databas present formal result literatur prove theoret lower bound space cost support approxim histor quantil queri bound reveal fundament differ answer quantil queri past present time propos structur find approxim histor quantil show consum space lower bound squarelogarithm factor extens experi demonstr practic techniqu perform predict theori quantil return remark accur theoret precis guarante concis represent rang queri top dk queri tempor data concis represent rang queri semant rank queri probabilist data expect rank improv transact dtime dbm perform function revers furthest neighbor spatial databas abstract set point queri point revers furthest neighbor rfn queri fetch set point furthest neighbor point monochromat rfn mrfn queri interest version rfn queri bichromat revers furthest neighbor brfn queri set point queri set queri point brfn queri fetch set point furthest neighbor point rfn queri interest applic spatial databas instanc larg residenti databas set potenti site build chemic plant complex construct site select maximum number revers furthest neighbor instanc brfn queri paper present challeng queri propos effici tree base algorithm monochromat bichromat version rfn queri analyz properti rfn queri differenti wide studi revers nearest neighbor queri enabl design algorithm approach take advantag furthest voronoi diagram convex hull data set mrfn case queri set brfn case brfn queri extend analysi situat larg size disk resid experi synthet real data set confirm effici scalabl propos algorithm bruteforc search base approach rank distribut probabilist data rank queri essenti tool process larg amount probabilist data encod exponenti determinist instanc applic uncertainti fuzzi inform aris data collect multipl sourc distribut network locat distribut sensor field imprecis measur multipl scientif institut inconsist scientif data due network delai econom cost commun larg amount data network fundament problem scenario retriev global top tupl distribut site minimum commun cost wellfound notion expect rank tupl world basi rank work design commun comput effici algorithm retriev top tupl smallest rank distribut site extens experi synthet real data set confirm effici superior algorithm straightforward approach forward data server robust approxim aggreg sensor data manag system emerg area sensor base system signific challeng develop scalabl fault toler method extract inform data sensor collect approach data manag problem sensor databas system user perform aggreg queri min count avg read sensor network addit advanc queri frequenc count quantil estim support due energi limit sensor base network central data collect gener impract system network aggreg reduc network traffic aggreg strategi remain bandwidth intens combin fault toler multi path rout method environ avoid expens investig approxim network aggreg small sketch present duplic insensit sketch techniqu implement effici small sensor devic limit hardwar support analyz perform accuraci final present experiment evalu valid effect method small synops group dby queri verif outsourc data stream preliminari releas articl accept acm transact databas system definit version product acm releas supersed version effici process top dk queri uncertain databas abstract work introduc polynomi time algorithm process top queri uncertain databas gener adopt model relat relat consist number tupl tupl randomli instanti tupl altern result significantli improv algorithm top queri process uncertain databas term run time memori usag focus singl altern case algorithm order magnitud faster random synops queri assur data stream due overwhelm flow inform data stream applic compani acquir resourc deploi data stream manag system dsm choos altern outsourc data stream desir comput parti data outsourc remot comput intrins rais issu trust make outsourc queri assur data stream problem import practic implic set continu group sum queri process remot untrust server client limit process capabl observ stream server regist queri server dsm receiv result request client verifi integr result significantli fewer resourc evalu queri local goal propos probabilist verif algorithm select aggreg group queri constant space irrespect result set size low updat cost stream element arbitrarili small probabl failur gener algorithm toler number erron group detect order support semant load shed server discuss hard support random load shed final implement techniqu perform empir evalu live network traffic find frequent item probabilist data comput statist inform probabilist data attract lot attent recent data gener wide rang data sourc inher fuzzi uncertain paper studi import statist queri probabilist data find frequent item straightforward approach identifi frequent item probabilist data set simpli comput expect frequenc item decid exce fraction expect size data set simpl definit miss import inform intern structur probabilist data interplai uncertain entiti propos definit base world semant wide adopt queri type uncertain data manag find item frequent randomli gener world approach natur lead studi rank frequent item base confid find frequent item probabilist data turn difficult propos exact algorithm offlin data run quadrat cubic time design sampl base algorithm stream data find approxim frequent item theoret guarante high probabl accuraci sampl scheme consum sublinear memori exhibit excel scalabl final verifi effect effici develop algorithm real synthet data set extens experiment evalu trip plan queri road network databas effici process top dk queri uncertain databas drelat proof dinfus stream enabl authent slide window queri stream time seri compress privaci paper studi trade off time seri compress partial inform hide fundament implic introduc uncertainti individu valu perturb specif perturb compress properti origin data detect filter reduc uncertainti make perturb similar origin data preserv structur data simultan make breach harder data compress fraction uncertainti remov true valu leak reveal perturb formal notion studi trade off real data develop practic scheme strike good balanc extend fly data hide stream environ character exploit refer local data stream applic paper investig approach process queri data stream applic show refer local characterist data stream exploit design superior flexibl data stream queri process techniqu identifi refer local popular long time scale tempor correl shorter time scale eleg mathemat model shown precis quantifi degre sourc local analyz impact local awar achiev perform gain tradit algorithm applic max subset approxim slide window join approxim count estim comprehens experiment studi compar exist algorithm local awar algorithm number real dataset result valid us effici approach trip plan queri spatial databas paper discuss type queri spatial databas call trip plan queri tpq set point interest space point belong specif categori start point destin tpq retriev trip start pass point categori end driver travel boston provid stop ga station bank post offic goal provid rout term distanc traffic road condit difficulti queri li exist multipl choic categori paper studi fast approxim algorithm tpq metric space provid number approxim algorithm approxim ratio depend number categori maximum number point robust aggreg sensor network emerg area sensor base system signific challeng develop scalabl fault toler method extract inform data sensor collect approach data manag problem sensor databas system user perform aggreg queri read sensor network due power rang constraint central approach gener impract system network aggreg reduc network traffic aggreg strategi bandwidthintens combin fault toler multi path rout method environ order avoid expens investig approxim network aggreg small sketch survei robust scalabl method comput duplic sensit aggreg spatio dtempor aggreg sketch spatio tempor applic requir retriev summar inform move object lie queri region queri interv number mobil user cover cell traffic volum district exist solut distinct count problem object remain queri region timestamp queri interv count multipl time result paper solv problem integr spatio tempor index sketch tradition approxim queri process propos techniqu appli reduc space requir convent spatiotempor data mine spatio tempor associ rule approxim aggreg techniqu sensor databas approxim aggreg techniqu sensor databas build logic view websit wiccap data model map physic websit logic view inform sourc www larg amount data organ interest valu import facil enabl user extract inform interest simpl effect manner inform web sourc extract automat user interest extract inform requir depth knowledg relev technolog extract process slow tediou difficult ordinari user propos wiccap data model xml data model map web inform sourc commonli perceiv logic model base data model ordinari user extract inform easili effici acceler creation data model defin formal process creat data model implement softwar tool facilit autom process produc wiccap data model visual tool build logic data model websit inform sourc www larg amount data organ interest valu import facil enabl user extract inform interest simpl effect manner propos wiccap data model xml data model map web inform sourc commonli perceiv logic model inform extract automat user interest acceler creation data model implement visual tool call map wizard facilit autom process produc wiccap data model tool time requir construct logic data model websit significantli reduc inform concierg web www inform collect collag program wiccap system softwar system gener logic view web resourc extract desir inform form structur document design enabl peopl obtain inform interest simpl effect manner make inform www access applic afford autom inter oper web awar servic kei factor make system practic tool autom facilit process construct logic represent websit identifi defin inform interest retriev paper present design wiccap system main compon map wizard network extract agent 