toqbianditssuitabilityasaqueryprocessingfacilityformobileusersisformally discussedinsection5.section6describesandevaluatesthreealgorithmsforgen-
|
|
- Cordelia York
- 8 years ago
- Views:
Transcription
1 SupportingMobileDatabaseAccessthrough QuerybyIcons ckluweracademicpublishers,boston.manufacturedinthenetherlands.,,1{23() ANTONIOMASSARI Dipart.diInformaticaeSistemistica,UniversityofRome"LaSapienza",00198-Roma,Italy SUSANWEISSMAN,PANOSK.CHRYSANTHIS DepartmentofComputerScience,UniversityofPittsburgh,Pittsburgh,PA15260,U.S.A. sincetheycanbemanipulatedwithouttyping.thefacilityrequiresnospecialknowledgeofthe queryprocessingfacilitythatsupportstheexplorationandqueryofdatabasesfromamobile computerbasedonthemanipulationoficons.iconsareparticularlysuitableformobilecomputing Abstract.Inthispaper,wepresentboththetheoreticalframeworkandaprototypeofa locationorthecontentoftheremotedatabasenorunderstandingofthedetailsofthedatabase schema.itsiconicquerylanguageinvolvesnopathspecicationincomposingaquery.the Keywords:MobileQueryProcessing,IconicQueryLanguage,MobileComputing,Semantic communication. queryfacilityprovidesmetaquerytoolsthatassistintheformulationofcompletequeriesinan Distance,SemanticandObjectModel,PathComputations eectivelycopewiththeinherentlimitationsinmemoryandbatterylifeonthemobilecomputer, disconnectionsandrestrictedcommunicationbandwidth,andthehighmonetarycostofwireless theremotedatabase.bynotrequiringconstantaccessandcachingoftheactualdata,itisableto incrementalmanneronthemobilecomputerandwithoutinvolvingaccesstotheactualdatain fectedthewaythatwecomputebut,moresignicantly,theyarechangingtheway 1.Introduction Advancesincomputerandwirelesscommunicationtechnologieshavenotonlyaf- weliveanddobusiness.forexample,mobileusers,sayahospitalparamedicunit capabilityofquicklylocatingandcontactingmedicalpersonnelnearesttotheaccidentsite.thatis,mobileusersbymeansofhand-heldcomputersequippedwitha wirelessinterface,shouldbeable(1)tocomposeadatabasequerywithminimum arrivingatanaccidentsite,needstoeasilyaccessthemedicalhistoryofthevic- ornoknowledgeofhowthedatabaseisstructuredandwhereitislocated,and(2) tocomposethequerywithafewkeyselectionsandminimumtyping[3]. tims,regardlessofthelocationandformoftheinformation.theyalsoneedthe calledquerybyicons(qbi),considerstheinherentlimitationsinmemoryand facilitysuitableformobiledatabaseapplications.thequeryprocessingfacility, batterypoweronthemobilecomputer,disconnectionsofthemobilecomputerfor substantialperiods,restrictedcommunicationbandwidth,andhighmonetarycost ofwirelesscommunication[4,15]. Motivatedbytheserequirements,wepresentinthispaperaqueryprocessing
2 Aniconicvisuallanguageinterface,whichallowsausertocomposeadatabase 2ThesalientfeaturesofQBIarethefollowing: Asemanticdatamodelthatcaptureslocallywithinthemobilecomputermost icallygeneratednaturallanguage. querybymanipulatingiconsusingapointingdevicelikealight-penonahandheldpen-computer.bothstructuralinformationandconstraintsarevisualized focalobject,asclassesofobjectsexpressedasgeneralizedattributesofthefocal ofsimplerepresentationstructures.thatis,auserisnotrequiredtohave oftheaspectsofthedatabasestructureswhilepresentingtheuserwithaset ofthedatabaseschema.auserperceivesthewholedatabasefromanysingle anyspecialknowledgeofthecontentoftheunderlyingdatabasenorthedetails whereastheimplicitambiguityoficonicrepresentationisresolvedbyautomat- Intensionalormetaquerytoolsthatassistintheformulationofacompletequery object.generalizedattributesencapsulateandhidefromtheuserthedetails Whileaniconicinterfaceallowsfastinteractions(fasterthantyping)evenwhen accessingactualdataintheremotedatabasetomaterializeintermediatesteps. Dataareaccessedandtransmittedbacktothemobilecomputeronlywhena completequeryismaterialized. duringdisconnections.aqueryisformulatedinanincrementalmannerwithout ofspecifyingaquery. theuserismoving,queryformulationusingintensionaldataosetstheexpense actualdatabase.frequentcommunicationresultsinslowerresponsetimeduetothe limitedbandwidthofwirelesslinks,aswellasconstantdepletionofthecomputer's andlimitationsoffrequentwirelesscommunicationthatisinherenttoextensional battery.therefore,userscanplaninadvancetobedisconnectedfromthenetwork browsingsystems,e.g.,[27,29].metadatacanbecachedonthediskonthemobile computersinceitsdenitionchangesrarelyanditssizeissmallcomparedtothe inordertosaveenergyandreducecommunicationcostswhileactivelyexploringthe databaseviaintensionalinformationonthemobilecomputer.inaddition,users cancontinuewiththeformulationoftheirqueriesonthemobilecomputereven andstationaryhoststoqueryandexplorealargedistributeddatabasemanaged whenthecomputerisaccidentlydisconnected. byanumberofserversonstationaryhosts.inthenextsection,theconceptional modelofqbithatdenesthemobileuser'sperceptionofadatabaseisdiscussed. Section3describesthecomponentsofaQBIprototypeandillustratesitsfunctionalitythroughitsusetoqueryamedicaldatabase.QBI'stheoreticalframeworkis toqbianditssuitabilityasaqueryprocessingfacilityformobileusersisformally discussedinsection5.section6describesandevaluatesthreealgorithmsforgen- Inamobiledatabaseenvironment,weenvisionQBIbeingusedonbothmobile presentedinsection4whereasthenotionofgeneralizedattributeswhichiscentral eratinggeneralizedattributesonamobilecomputer.thepaperconcludeswitha discussiononrelatedworkinsection7andfutureworkinsection8.
3 2.QBI'sConceptualDataModel InQBI,theconceptsofclassofobjectsandattributeofaclassexclusivelyformthe externalrepresentationofthedatabasestructureduetotheirnaturalsimplicity. Usersarepresentedwithadatabaseabstractioncalledcompleteobjects[25],i.e., 3 relationaldatabases[23].specically,auserperceivestheunderlyingdatabaseas completelyencapsulatedobjects,similartotheuniversalrelationabstractionin throughanexample.assumetheunderlyingdatabaseschemadepictedinfigure1 asetofclasses,eachhavingseveralpropertiescalledgeneralizedattributes(ga). usingthebinarygraphmodel[9,25],whichalsoformsqbi'stheoreticalframework(seesection4).heretherectanglesdenoteobjectclassesandovalsconvey propertyofanentity,agaexpressesagenericpropertyofaclass. jectswithineachfocalobject.thatis,otherobjectclassesareviewedasgasof underlyingdatabasefromitsownviewpoint.letusillustratetheconceptofgas thefocalobject.thus,inqbi,eachfocalobjectprovidesaviewofthewhole InthesamewaythatanattributeintheERmodel[10]representsanelementary theinteractionamongclasses. Further,GAsencapsulatebothimplicitandexplicitrelationshipsamongtheob- OfthethreeGAsoftheclasspersonshowninFigure1,considerattributeGA3 AQBIuserobservesthattheunderlyingdatabasecontainsthesameobjectclasses tirestructureofthedatabasebymeansofthegasofanyoftheseobjectclasses. showninfigure1,namely,person,car,city,andhospital,butviewstheen- Figure1.ExamplesofGeneralizedAttributes sonscorrespondingto,"allthepeoplelivinginthesamecitywherethe Thatis,fromtheviewofahospitalobjectclass,thisGAisasubsetofper- perceivesthatahospitalislocatedinacityandisanattributeofperson.a generalizedattributewithsimilarmeaningexistsfromtheviewpointofhospital. whosevalueisasubsetoftheobjectclasshospital.byobservingga3,theuser 1,1 car owns GA2 0,n GA1 GA3 GA1: Birthdate of the person GA2: Set of cars owned by the person GA3: Set of hospitals located in the same city where the person lives 1,1 0,n born 1,1 lives 0,n date city 0,n located 1,1 hospital
4 4hospitalislocated".Also,thesameinformation,couldbeobtainedbyobservingtheGAsofcity. amedicaldatabasethatincludesradiologicaldatafromamobilecomputer. 3.QBIPrototype Inthissection,wewilldescribeourQBIprototypeandhowitcanbeusedtoquery isonly0.5mbyteswhereasitstoreslessthan100kbytesofintensionalinformation abouttheradiologicalmedicaldatabase. toolkitxvt,anditcurrentlyrunningonncrsystem3125pen-topcomputerswith PenDOSandMS-WindowsforPenComputing.ThesizeoftheQBIprototypeitself TheQBIprototypeiswritteninCfortheMS-Windowsenvironmentusingthe Figure2.ThearchitectureofQBI thedatabaseonastationaryhost,aswellasmanagingthesporadicupdatesto consistsoffourmodules:thepresentationmanagerwhichisresponsibleforall themetadataandstatisticsabouttheunderlyingdatabase.gasarecomputedon demandbecausegivenanobjectclass,potentially,thereareaninnitenumberof interactionswiththeuser;thequerymanagerwhichsupportsthespecication AccessManagerwhichisresponsibleforanyremoteaccesstotheactualdatain GAsassociatedwithanobjectclassandonlyasmallfractionofthemareusefulin ofqueries;gaevaluatorwhichcomputesthegeneralizedattributesanddatabase TheoverallarchitectureoftheQBIprototypeisdiagrammedinFigure2and thegaevaluator,issupportedbytwodatabases,namely,theintensionaldatabase (IDB)andStatisticaldatabase(SDB).TheIDBcontainsallthemetadataand thevisualdatafortheiconicrepresentation.whereassdbcontainsstatistical informationontheinstancesinthedatabaseusedfortheevaluationofthegas. theconstructionofaparticularquery. Theexecutionofthepresentationmanageraswellasofthequerymanagerand QBI User Requests Queries Presentation Manager Query Manager Database Access Manager Network IDB GA Evaluator SDB
5 Thepresentationmanagerstructurestheinteractionswithauseraroundthreewindows,eachdedicatedtoaspecicaspectinthespecicationsofaquery.Thethree windowscomposingtheqbiinterfacearereferredtoastheworkspacewindow, thequerywindowandthebrowserwindow QBI'sIconicVisualLanguageInterface WorkspaceWindow WhentheQBIapplicationstarts,theuserisaskedtoselectthedatabasetobe storedqueries(seefigure3). tiveclassesthatareactuallystoredinthedatabaseandderivedclassesrepresenting Workspacewindowappearsanddisplaysasetoficonscorrespondingtobothprimi- consideredforquerying,inourexamplearadiologicaldatabase.inresponse,the shapeofaniconcanappearat,suchasthepatienticon,orasastackofshapes, theimageisalabelthatallowsforeasyidentication.afullnaturallanguage associatedwithaspecicgeometricoutline,similartoajigsawpuzzlepiece.the sentencedescriptionisalsoprovidedthatcanbereadbypointingattheicon.this descriptionisautomaticallygeneratedbasedonthemethoddescribedin[8]andis essentialfordisambiguatingthemeaningbetweenvariousicons.eventheshapeof objectclassesthatareallowedtobecombinedinaselectionconditionofaqueryare theiconconveysinformation.whenformingaquery,iconsrepresentingcompatible Everyiconhasanimageconveyingametaphoricalmeaningfortheclass.Below representsagroupofhospitals QuerySpaceWindow PointingataniconintheWorkspacecorrespondstoselectingitsobjectclassfora suchasthehospitaliconfoundatthebottomofthebrowserwindowinfigure queryviatheactivationofthequeryspace.iftheclassiconforpersonispicked 3.Thisstackedrepresentationtellstheuserhowmanyinstancesoftheobjectclass fromtheworkspace,thequeryspaceshownatthebottomleftoffigure3will canbereferredtowiththisoneicon.thehospitaliconappearsstackedsinceit ConditionsSpace:Thisspaceontheleftsideofthequerywindowallowstheuser tocomposeaquerybasedontheselect-projectparadigm: becomevisibletotheuser.thereareseveralpartstothiswindowthatallowauser tobuildbothgasconstitutingtheatomsofaselectionconditionandthe
6 6 Figure3.TheQBIInterface Figure4.AQuery
7 ShowSpace:Iconscanalsobearrangedinthesectionontherightcalledthe conditionitself.atomscanbecombinedtogether,accordingtoapositional convention,toformthebooleanexpressionrepresentingtheselectioncondition. ShowSpacewhichisusedforspecifyingtheprojection.Theseiconsrepresent theinformationtheuserchoosestoviewintheoutputresult.aninitialga7 DescriptionSpace:Thisspacecontainsanaturallanguagedescriptionoftheclass beingdened.thedescriptionisautomaticallygeneratedanddynamically andcorrespondstotheattributesthatwouldappearinanequivalententity setisdisplayed(bydefault)intheshowspacewhenaclassiconispicked ofperson. updatedwhenevertheselectionconditionschange. classpersonare:nameoftheperson,sexofthepersonandbirthdate Relationshiprepresentationofthedatabase.Forexample,theinitialGAsofthe TheBrowserWindowistheinterfaceoftheGAgeneratorthatallowsauserto TheBrowserWindowandMetaquerying class,howmeaningfulaparticulargaisfortheobjectclass.theadditionalsetof semanticdistanceorweightthatcharacterizes,fromtheviewpointofanobject perceptionofthemostmeaningfulattributesofperson.thebrowserwindowof notincludedintheinitialgasetdisplayedbypressingthebuttonlabeledmore attributes.withagivenobjectclass,itsgasgenerationiscontrolledbya windowactivatedbyselectingtheiconperson,ausercanseeadditionalgas GAsissortedbytheirsemanticdistancesothatthemostmeaningfulGAsareshown rst.thus,byobservingthetopofthelistofgastheusercanhaveanimmediate exploreadatabasebycontrollingthegenerationofgas.fromthequeryspace andshowspaceswithinthequerywindowwhenformingaquery. metaqueryoperatorspermitthespecicationoflterconditionsonthegaset. Hence,auserinterestedinverydistantpropertiesoftheclasspersoncaneasily environment,asetofmetaquerytoolsareprovidedwithinthebrowser.these suchasthisonecanthenbedraggedfromthebrowserwindowintotheconditions Figure3showstheadditionalGACitywherepersonXlives.AdditionalGAs desiresgaswhichareassociatedwithaspecicobjectclass.iftheuserisinterested inallthegasthattalkaboutcity,theiconforcityfromtheworkspacewindowcan smallergaset.forexample,oneusefulmetaqueryoperatorisusedwhenauser exploretheseproperties,byrestrictingthesearchofthedesiredgaswithina bemovedintothetalkaboutspaceofthebrowser(seefigure3).inparticular Toempowertheuserwiththeabilitytocontrolhis/herviewofthedatabase ofaparticularobjectclassand,(3)talkabout,don'ttalkaboutselectsgasthatare itispossibletoexpressthefollowingmetaqueryconditions:(1)single,printable, orkeyselectsonlysinglevalued,printablegas,orkeygasusedtoidentifyan instanceofaclass,respectively,(2)typeselectsallthegasthatrepresentasubset
8 8associatedornotassociatedwithaspeciedclass.Allthemetaqueryconditions 3.2.QueryExamples Letusrevisitthehospitalparamedicunitexamplementionedintheintroduction. arecombinedinaconjunctiveexpressionbydefault. ofthisqueryisasubsetoftheclassdoctorthatcanbesavedasaderivedclass. closeproximityofthehospitalisnotied.forthistypeofinformation,weneedto andcitywheredoctorxworkswiththeconnectiveisequalto. determinethesetofdoctorslivinginthesamecityinwhichtheywork.theresult Asapatientisrushedtothemostappropriatehospital,aspecialistlivingwithin ThisGAisdraggedbytheuserfromtheBrowserwindowintotheconditionspace Inordertobuildthederivedclassitisnecessarytospecifytheselectioncondition Cond:Thecitywherethedoctorworksisequaltothecitywherethedoctorlives. CondcanbespeciedbyconnectingthetwoGAs:Citywheredoctorxlives ofthequerywindow.asfarasthesecondgaisconcerned,theuserneedsto performametaqueryonthegasetofdoctorbydraggingtheiconcityintothe \best"connectionbetweendoctorandcityanditcoincideswiththegatheuser TypespaceoftheBrowserwindow.TherstGAshowninthelistrepresentsthe waslookingfor,thatis:citywhereahospitalislocated.suchahospitalisthe hospitalwheredoctorxworks. TherstGAisimmediatelyfoundbyscrollingthelistintheBrowserwindow. thisoperationtobeperformed.oncethetwogashavebeenattachedtogether,a \attached"together;sincetheyhavethesametype(thatofcity),theirshapesallow connectivetheuserendstheselectionpartofher/hisquery.inthedescriptionspace dialogueboxcontainingasetofvalidconnectivesappears.bychoosingtheequality doctorbecausetheyarepartoftheinitialgasetofdoctorandmustbealready asentenceexplainingtheselectionqueryisautomaticallyadded(figure4).for intheshowspace.inordertoknowwhichcitythehospitalanddoctorarelocated, theprojection,theuserdoesnothavetopickthename,sexandbirthdateofthe ThesecondGAisthendraggedintotheconditionspaceandthetwoGAsare isassignedaniconwhichbecomespartoftheiconsetcontainedintheworkspace windowandcanbeusedasifitwereaprimitiveone.theimageofthisiconwillbe thisquery,alltheuserhastodoisdragthisnewicontoaspecialsystemicon thatofadoctorsincethisgaisasubsetoftheobjectclassdoctor.tomaterialize choosesalabel,sayluckydrs,forthenewlycreatedderivedclass,thenewclass itcanbestoredintheintensionaldatabase.intheaboveexample,aftertheuser theusersimplydragstheappropriate,i.e.,city,iconfromthebrowserintothe querytotheremotedatabase,requestingforitsexecution,andfetchingthequery calledthe\printer".thisoperationcorrespondstoforwardingthecorresponding Theresultofaqueryconstitutesaderivedclassofthepickedclassandassuch resultfordisplay.
9 mentasqueriesrequireverylittletyping.also,onlyasmallscreenisrequired helpfulforuserswithalimitedknowledgeofdatabaselanguages.inaddition,intensionaldataandmetaquerytoolsareprovidedtouserstoallowthemtoformulate sincequeriesdonotrequireanyformofpathspecication.theunderlying,formal schemaishiddenandfeedbackintheformofnaturallanguageandshapesisvery Asseenwiththisexample,QBIisveryusefulinamobilecomputingenviron- 9 4.QBI'sTheoreticalFramework queriesevenwhenthecomputerisdisconnected.inthefollowingsection,wewill Internally,QBIusestheBinaryGraphModel(BGM)[9,25],asemanticdatamodel, theinternalalgorithmsforgagenerationwereimprovedforamobileenvironment. generalizedattributesandsemanticdistance.infurthersections,wewillshowhow formallydenethesemanticdatamodelusedbyqbi,aswellastheconceptsof forcapturingmostoftheaspectsofthestructureoftheremotedatabase.themajor constructsofthismodelare:theclassofobjects,thebinaryrelationshipamong setnofnodesconsistsofclassnodesncrepresentingclassesofobjectsandrole nodesnrrepresentingrelationshipsbetweentwoclasses.classnodescanbeeither schemacanbeexpressedasalabeledgraphcalledtyped-graph. Denition4.1-TypedGraph.Atyped-graphg(N;E)isalabeledmultigraph.The classes,theisarelationshipbetweenaclassanditssuperclass,andcardinality printableornonprintabledependingonwhethertheyrepresentdomainsofvalues orabstractclasses.anedgeinecanonlylinkaclassnodetoarolenodeandis constraintsfortheparticipationofclassinstancesintotherelationships.abgm associatedwithauniquelabell.eachrolenodehasadegreeequaltotwo. representedbygandcwhereasaninstanceofadatabase(extension)arerepresented isasetofconstraints,andmisaninterpretation.theschemaofadatabaseis thetwonodes.eachrolenodewillhaveexactlytwoadjacentclassnodes.when bythenotionofinterpretation. theadjacentclassnodesarecoincidentwesaythattherolenodeisreexive.in thiscase,labelsonedgesareusefulfordisambiguatingthetwoedges. ABGMdatabaseisdenedasatriple<g;c;m>,wheregisatyped-graph,c Aclassnodeissaidtobeadjacenttoarolenodeifthereisanedgeconnecting Denition4.2-Interpretation.Letgbeatyped-graph.Aninterpretationfor gisafunctionmmappingeachclassnodenc2nctoasetm(nc)ofobjects wherelbl1;lbl2arefunctionsreturningthelabelsofthetwoedgesconnectedtonr adjacentclassnodeofnr. andeachrolenodenr2nrtoasetm(nr)ofpairs<lbl1(nr):x1;lbl2(nr):x2>, (lbl1;lbl2:nr!l)and<x1;x2>2m(nc1)m(nc2)wherenc1andnc2arethe
10 10Thatis,aninterpretationspeciesthevalidcombinationsofvaluesfromthe underlyingclasses.thesetofconstraintsonthedatabasereferredtointhispaper aretheminimum(atleast)andmaximum(atmost)cardinalityconstraints, andthesubclass-superclassrelationshipconstraint(isa). Denition4.3-Constraints.Thesetccontains:(1)ATLEAST(k;nc1;nr)speciesthataninstanceofclassnodenc1canparticipateinatleastkinterpretation involvingtheadjacentrolenodenr(2)atmost(k;nc1;nr)speciesthataninstanceofclassnodenc1canparticipateinatmostkinterpretationinvolvingthe adjacentrolenodenr.(3)isa(n^c;nc)speciesthattheclassn^cisasubsetofthe classnc(i.e.,m(n^c)m(nc)).therolenodesconnectedtoncareconsideredas alsobeingconnectedton^c,i.e.,n^cinheritstheedgesofnc. Currently,weassumesingleinheritance,henceeachclassnodehastobelongto oneandonlyoneclasshierarchy.inordertofacilitatetypecheckingofquery expressions,wedenethenotionoftypeofaclassasaclasshierarchy.thatis,the classnodesbelongingtoaclasshierarchyhavethesametype.notethatifnc1and nc2havedierenttypes,theirinterpretationsaredisjoint. person string date hospital car city name name model name born lives located owns 0,n 1,1 1,1 0,n 1,1 0,1 0,1 0,1 0,1 1,1 0,n 1,1 1,1 0,n 1,1 1,1 patient disease disease heart name 0,1 1,1 had 0,n 0,n Figure5.ATypedGraph Figure5showsanexampleofatypedgraphwhichisafragmentofthemedical databaseusedintheprevioussection.therectangularboxesrepresentclassnodes (theprintableonesaregrayed)whiletheovalsrepresentrolenodes.nolabelonan edgeisshownsincetherearenoreexiverolenodesthatneedtobedisambiguated. Theannotations(m,n)onedgesrepresent(ATLEAST,ATMOST)cardinalityconstraints.ISAconstraintsaredenotedbyathickarrowfromasubclassnodetoits superclass.
11 Asmentionedabove,theconceptofGAinQBIrepresentsthewayinwhichauser perceivestherelationshipsamongobjects.internally,agaisstrictlyrelatedto 5.GeneralizedAttributes theconceptofpathinatyped-graphcapturingthedatabaseschema,whereapath 11 isasequenceofadjacentclassandrolenodesalwaysstartingandendingwitha withfirst(p)andlast(p)respectively. classnode. Denition5.1-Path.LetGbeatyped-graph.AstepsonGisthetriple <class1(s),role(s),class2(s)>whereclass1(s)=nc1;class2(s)=nc22ncare adjacenttorole(s)=nr2nr.apathpongisthesequences1;s2;skofsteps classnodeofapathp,i.e.class1(s1(p))andclass2(slength(p)(p)),willbedenoted ongsuchthat,fori=1k?1,class2(si)=class1(si+1).therstandthelast endingin^ncdenesagaofncasafunctionmappingeachinstancexofnconto Denition5.2-GeneralizedAttribute.LetGbeatyped-graph,ncaclassnodeofG asetofinstancesof^nc. andpapathongsuchthatfirst(p)=nc;thegaoftheclassnodencassociatedto pisafunctionp:m(first(p))!}(m(last(p)))mappingeveryobjectx02m(nc) toasubsetofobjectsofthelastclassnodeofp,m(last(p)). Giventwoclassnodesnc(pickedclassnode)and^nc,apathpstartinginncand AGAcanbeeithersinglevaluedormultivalueddependingonthecardinalityconstraintsoftherolenodesinvolvedinthepath.SinceaGAisafunctionpreturning typeoflast(p). asetofobjectsbelongingtom(last(p))wewillsaythatphasatypethatisthe onlythosegasthatare\meaningfulenough"forthespecicationofaquery.a representusefulproperties,e.g.,peoplelivinginthesamecitywherethe similarnotiontothatofsemanticdistancefunctioninqbiisthenotionofsemantic withinnitelylong(cyclic)paths,qbidenesasemanticdistancefunctionon personlives.asaconsequencethesetofpossiblegasassociatedtoaclass correspondingga.anitesetofgasofanobjectisconstructedbyconsidering pathswhichreturnsavalueforeachpathrepresentingthemeaningfulnessofthe nodeisinnite.sincenotallpathsareequallymeaningful,andinordertocope Apathinatyped-graphcanbecyclic.Cyclicpathsareallowedsincetheycan lengthinthegenerationofcompletequeriesfromincompletepathexpressions[17]. Denition5.3-SemanticsDistanceFunction.LetpbeaGAofaclassnodenc. ThefunctionSemd:?(nc)!<mapsptoarealvalueSemd(p)thatrepresents howmuchpissemanticallydistantfromnc.
12 12Thesemanticdistanceisexpressedintermsofvariousaspectsofthestructureof nameofacar,"canbediscarded. lessmeaningfulgas,suchas"allthepersonsthathaveanameequaltothemodel thegasuchasthelengthofthepath,thenumberofcycles,inclusion/exclusion conveyssomeinformation.bytakingentropyintoconsideration,alargenumberof TheentropyclearlycapturesthefactthatauserconsidersaGAonlyiftheGA ofspecicpaths,cardinalityconstraintsaswellasstatisticalinformationonthe tropyofagawhichisthemeasureofuncertaintyintheinformationtheory[13]. underlyingdatabase.thestatisticalinformationcanbeusedtocomputetheen- providedthatsemdismonotonicallyincreasing. GAsetofnc,withrespecttoSemd,willbedeterminedby: GivenasemanticdistancefunctionSemdandathresholdvalue2<,thenite GAp.Letthenewstepbeingaddedtopbee=<u,v,w>(i.e.,p0=p[e). semanticdistanceofanewgap0resultingfromaddinganewsteptoanexisting ThefollowingfunctionwasimplementedintheQBIprototypetocomputethe?Semd(nc)=fp2?(nc)jSemd(p)<g Semdp0=(c1length(p0))+(c2numcycles(p0)+ (c3maxcardinality(p)atmostcardinality(e)-1)+npwc4+ lengthisdenedtobethenumberofrolenodesconnectorsusedbythepathand mostcardinality(e)),sodoesthesemanticdistanceassociatedwiththisga.the wherenpw,npuandeareagswith0or1value,andc1;c2;c3;c4;c5andc6 existinanewgapath.asaga'spathbecomeslonger(length(p0)orcyclic (numcycles(p0)),oritscardinalitydramaticallyincreases(maxcardinality(p)at- arepositiverealconstantsusedassemanticpenaltiesforpropertiesthatmay E(c5=avgcardinality(p0)?c5)+NPUc6 doesnotincludeanyclass-superclassconnectors(i.e.,``isa''connectors). lationshipbetweenncandanotherabstractobjectclass.thisobjectclasswill noinformationfortheuserthatisdirectlyobtainablefromthispath,thepathis penalizedbyassigningnpwto1.otherwisenpwis0. containsimpleattributesthathavenotbeendiscovered.however,becausethereis nodesis300andisequivalenttothepenaltyc2associatedwithaclassnodebeing amemberofacycle.theotherpenaltyconstantsarec1=200,c3=0:02,c4=20 Apaththatendswithaclassnodewwhichisnotprintable,representsare- andc5=100.allpenaltyconstantshavebeadjustedforbetterperformanceafter aseriesofexperimentswhilemakingsurethatthesemdremainsmonotonically increasing. IntheQBIprototype,thecurrentweightincreasec6forgoingthroughprintable
13 6.MobileGAGeneration Inamobileenvironment,QBI'smethodofqueryformulationanduseofintensional queryformulationisdoneprimarilythroughthemanipulationofgeneralizedattributes.asstatedabove,agapissimplyapathpinatyped-graphg.every timeastepisaddedtoanexistinggapath,anewgaisformedanditsasso- oftheenvironmentinwhichqbiisexecuting.ontheotherhand,thevaluefor denedasafunctionoftheavailablememory,theenergylevelandtheresponse ciatedreal-valuedsemanticdistancevalue(sd-value)iscomputedbythefunction alizationofcompletequeries.visualizationofthedatabaseaswellascosteective datalimitsthecostoffrequentwirelesscommunicationwithrespecttothemateri- 13 time: accuracytoexpressthemeaningfulnessofaga.however,thiscostisindependent thethresholdthatterminatesthegenerationofgascanbetunedtoconsiderthe capabilitiesofthesystem.inthecaseofthemobilecomputer,thethresholdis isclearlythemostcomputationallyexpensivepartofqbi.thecostofthesemanticdistancefunctiondependsonitscomplexitywhich,inturn,isameasureofits Semdp.ThegenerationofGAsisaninstanceofapathcomputation[2,16],and oryandenergylevel,varyovertime,thethresholddynamicallychangesaswell. whereciareuserdenedparameters.giventhatthersttwoparameters,freemem- =C1(freememory)+C2(energylevel)+C3(cpuspeed)+C4(resprequirements) usedinpathcomputations[2,16],ithasturnedouttobeunsuitableformobile depth-rstsearch(dfs)manner.althoughthisisalsotheapproachtraditionally operations.first,sinceitdoesnotgenerategassortedbasedontheirsd-value, anadditionalsortingphaseisrequiredforthepresentation.moreimportantly,the spendingasignicantamountoftimeingeneratinggaswithhighersd-values.for forthepossibilityofadynamicallysetthresholdtoterminatetheexecutionofthe GAevaluatorbeforethegenerationofGAsassociatedwithalowsd-valueandafter DFSstrategyisnotcompatiblewithadynamicallydenedthreshold.DFSallows TheinitialGAevaluatorinourQBIprototypetraversesthetyped-graphina thesereasons,wehaveexploredtwootheralternatives,thebest-rstsearch(best) andbreadth-rstsearch(bfs)basedgaevaluators. thegasfoundbasedonthesemanticdistancesassociatedwiththegas.ateach iteration,bestalwaysconsidersthegawiththeminimalsemanticvalue.the BEST-basedalgorithmuses:(1)asortedListtomaintainanorderamongtheGAs produced,(2)nc,thestartingclassnodechosenbytheuser,(3)apathpassociated witheverygapandatotalsemanticdistancesemd(p),and(4)thefunction, weight(u,w),whichcomputesthedistancebetweentwoclassnodesuandw,where Gisthegiventyped-graphandforeachstepe2G,e=<u,v,w>.Belowareshown thebasicstepsforthealgorithm. TheBESTalgorithmexploresthegivengraphGbymaintaininganorderamong
14 14 BEST(!=nc,p,,?(nc)) DoForeachstepe=<u,v,w>whereu=! Markeveryclassnodeinpofpasnewlyvisitedinordertodetectcycles. calculatesemd(p0)=semd(p)+weight(u,w) IfSemd(p0)Then WhiletheListisnotempty EndFor IftheListisnotempty p=firstoflist(list)!=last(p)(thepassociatedwithournewp)?(nc)=?(nc)[p p0=p[e SortedInsert(List,p0,p0,Semd(p0)) incrementalsortwithmanycomparisonswouldbeavoided.however,ithasturned moreoften,averysmallnumberofgaswouldbesortedeachtime,andbest's theclassnodesatalevelxwereexplored.although,sortingwouldbeperformed pathsthataretobeexpandedisnowdonebyaqueueinsteadofalist. levelareexploreddoesthealgorithmmoveontothenextlevel.alloftheitems usedbybestarealsousedbybfs,exceptmaintaininganorderamongthega typed-graphginalevel-by-levelfashion.onlywhenalltheclassnodesatagiven InaccordancetothebasicBFSalgorithm,theQueuewassortedeverytimeall TheBFSalgorithm,thesecondalternativeGAevaluator,exploresthegiven NPW,NPU,andEwhosevaluesaredependedonthetypeoftheclassnode,all ofgasendingatdierentlevelsalongdierentpaths.becauseofthebinaryags sortingphaseasdfs. 6.1.AdvantagesandDisadvantagesofDFS,BFS,andBEST outthatthisisnotenough,sincetherearenoorderguaranteesamongtheweights WithBESTandBFS,themainadvantageoverDFSiseciencyinndingmeaningfulGAs.BESTgeneratesGAsintheorderoftheirmeaningfulnesstotheuser basedontheirsemanticdistancefromthefocalobjectclassnc.withbfs,gas pathsproducedbypthatendatlevelxarenotguaranteedtohaveasmallersdvaluesthanallpathsfoundatlevelx+1.hence,bfsrequiresforanexplicit couldbesetbyafunctionthatdescribesthelimitationsofthemobileunit.with BEST,onlythemostmeaningfulGAswithrespecttotheselimitationswouldbe generated,anditdoesnotrequireaseparateexplicitsortingphase. withtheshortestpathsaregeneratedrst,anditishighlylikelythatthesepaths areverysemanticallymeaningfulfromtheviewpointofncduetothemonotonicallyincreasingpropertyofsemdalongapath.theorderingperformedbybest andbfsisusefulinamobileenvironmentsincethesemanticdistancethreshold
15 Ifthegraphisverydense,witheachnode!havingalargedegree,thenumberof combatthecostofinsertion,however,thisapproachwouldrequiremorespace.one comparisonswillbeveryhigh.anadditionalindexingstructurecouldbeusedto theformofasortedlistorqueue.intheworstcase,eachelementaddedtothe Listmustbecomparedtoalltheotherelementsbeforendingitscorrectlocation. However,unlikeDFS,bothBESTandBFSmaintaincomplexdatastructuresin 15 incrementalsortingoneachinsertionofanewgathewaybestdoes. advantagebfshasinthisregardisthatitdoesnotneedtoperformanexpensive fromthefocalpointobjectncarebeingexpandedconcurrentlyandbothreachthe requiresamarkerassociatedwiththenodew.iftheclassnodewisalreadya endingclassnodewtobecomeamemberofacycle.detectionofthisindfsonly partofthepathpbeforetheadditionofe,thenthenodewillbemarked.hence, marked.thispreventsthedetectionofafalsecyclewhenevertwoseparateroutes anotherfactorthatincreasestheexecutiontimeofbestandbfswhencompared todfs,isthemethodneededforcycledetection.inordertodetectifthenodewis partofacycle,alltheclassnodesofthepathpcurrentlybeingexpandedmustbe Inaddition,addingastepe=<u,v,w>toanexistingpathpmaycausethe samenode.however,thecostforthisisonetraversalofthepathpeverytime disadvantages,furtherinvestigationisnecessaryinordertodeterminewhichoneis Totheuser,thetwomostimportantcriteriaareresponse-timeandqualityofGAs i.e.,thesemanticdistancesofthegasproduced.thesearetwoimportantcriteria themostusefulwithrespecttoamobilecomputingenvironment. 6.2.EvaluatingtheGAGenerationMethods astepisadded.sinceeachofthesealgorithmshavecomparableadvantagesand indetermininghowwelleachofalgorithmsproposedforthegaevaluatorperform. However,withinamobileenvironment,theGAevaluatorshouldoperatewithout threeimplementationsunderadynamicallychangingthreshold.theseexperiments evaluated. regardstothequalityofattributesfoundbyeach.thesecondtestevaluatesthe discussedintheprevioussections.thersttestcomparesthethreealgorithmswith depletingalargeamountofthemobileunit'sresources.themethodproposed notionofadynamicallychangingthreshold.howeectiveachangingthreshold performswithrespecttondingmeaningfulattributesandresponse-timemustbe aboveforcontrollingtheamountofresourcesusedbythegaevaluatoristhe Test1:MeaningfulAttributeTest(MAT) weredoneusinganintel486dx,66mhzpcwith16mbytesram. TwotestsweredoneinordertoevaluatethethreeGAGenerationMethods forthedoctorobjectclassintheradiologicaldatabaseoftheqbiprototype. Task:Eachalgorithmwasgiventhetaskofndingagivennumber(X)ofGAs
16 16Parameters:Thesemanticweightthresholdremainedaconstant1800.At Eachalgorithmwastimedfromthemomentitwasinvokeduntilitwasableto produceasortedlistoftherequirednumber(x)ofattributes.inaddition,to measurethequalityoftheattributesproducedbythealgorithm,acomparisonwas madetoseehowmanyoftheattributesproducedbythealgorithmmatchedthe rstxattributesfoundinsal. Figure6isagraphofthesetestruns. thisvalue,thesystemproducedasortedattributelist(sal)of947gas.for eachsuccessivetestrun,thenumberofgasrequired(x)wasincreasedby100. Figure6.MAT:Qualityofattributesfound Test2:DynamicThresholdTest(DTT) Eachalgorithmwastimedfromthemomentitwasinvokeduntilitwasableto cessivetestrun.itwasincrementedbyavalueof100foreachrun. distancethresholdforthedoctorobjectclassintheradiologicaldatabaseof Parameters:Thesemanticdistancethresholdwaschangedforeachsuc- theqbiprototype. Task:EachalgorithmwasrequiredtondalltheGAs,belowagivensemantic numberofattributeswithingthatarebelowthecorrespondingthreshold. produceasortedlistofattributesbelowthegivensemanticdistancethreshold. Figure7isagraphthatshowshowmanysecondsittookeachalgorithmtond attributesinthetyped-graphgwillhavedistancesbelow.inordertocompare allthegasbelowagiventhreshold.foreachthreshold,acertainnumberof Inthisgraph,theamountoftimerequiredbyeachalgorithmismappedtothe theresultsofmatwiththistest,anothergraphshowninfigure8wascreated. Number matching first X of SAL dfs bfs best Number of attributes required (X)
17 # seconds to find attributes <= threshold 17 mationtheywantedifthebrowserwindowalwaysdisplayedtherst450most agiventhresholdvalue ImplicationsofMATandDTT Supposethatinamobileapplicationuserswouldbeabletoobtainalltheinfor- Figure7.DTT:TimetondattributesbelowFigure8.DTT:Numberofattributesfound belowagiventhreshold 0.72secondsusingBFScorrespondstoathresholdof=1600.Ofcourse,these meaningfulattributesforthedoctorobjectclass.figure6showsthatbfsnds thetop450mostmeaningfulattributesbygeneratingatotalof600attributesfrom thetyped-graphg.fromfigure7,itisapparentthatbfsndsthese600attributeswithin0.72seconds.also,fromfigure8,nding600attributeswithin 0 atablethatcorrespondstotheanswersfordfs,bfs,andbestforthebrowser Windowtoshow400,600,and800ofthemostmeaningfulattributestotheuser. gurescouldhavebeencalculatedusinganyofthethreeimplementations.belowis numberattributesdfssec.bfssec.bestsec. semanticdistancewhiledfsperformstheworstatgeneratingmeaningfulgas ThisresultwasexpectedsinceBESTdoesnotrequireanadditionalexplicitsorting phase.bfs'sperformancewasbetweendfsandbest.thealgorithmdoesrequire MATandespeciallyFigure6,showthatBESTndsGAsinorderofincreasing aexplicitsortingphase,butbecauseitgeneratesthegasbylevels,itsperformance intermsofgeneratingmeaningfulgaswasbetterthandfs dfs bfs best Maximum semantic weight threshold # seconds to find attributes <= threshold dfs bfs best Number of attributes found
18 matelythesameamountoftimeuntilthesemanticweightthresholdreaches1400. Soonafterthispoint,DFSbeginstotaketheleastamountoftimeandBESTthe increases,moreofthetyped-graphisexplored.allthreealgorithmstakeapproxi- most.thisphenomenoncanbeexplainedbynoticingthatastheexplorationofthe timeeachalgorithmrequirestodoitswork.thishappensbecauseasthethreshold 18InFigure7,weobservethatasthethresholdchanges,sodoestheamountof generatedthathavethesamepropertiesand,therefore,approximatelythesame graphmovesfurtherawayfromthefocalpointobjectclassnc,moreattributesare tohaveapproximatelythesamesemanticmeaningfromtheviewpointofnc.since BESTmustmaintainasortedlist,everytimealargegroupofcloselyweightedGAs semanticdistance.theselarge,similargroupsofgasarefarenoughawayfromnc thresholdbyafunctiondependingonhowmuchmemory,processing,batterypower, alargenumberoftraversalsandaccountsforthedecreaseinbest'sperformance areproduced,theymustbeplacedintheirproperpositioninthelist.thisrequires Withinamobileenvironment,themobilecomputercouldsetthesemanticdistance AnIntegratedMobileGAEvaluator asthethresholdincreases. littlevaluetotheuser.bestcangeneratethemostmeaningfulgaswithrespectto theselimitations.althoughthebestalgorithmismorecompatiblewithamobile environment,itsresponse-timesuersasthethresholdisdynamicallyincreased. tryingtoreserveitsresourcesandnotwastetimeorenergyndinggasthatareof inqbiwitheitherthebestorbfsalgorithms,ourexperimentationclearly tor.therefore,weproposetocombinethesetechniquesintooneintegratedmobile anddelaythehostandusercanaord.therefore,themobilecomputerwouldbe onthemobilecomputer.bestwillbethedefaultalgorithmofthemobilega GAevaluator.Sinceeachalgorithmonlyrequiresafewkilobytes,thisproposedintegratedevaluatorwillnotrequireadramaticincreaseintheamountofcodekept longasalowsemanticthresholdiscomputed,thebestalgorithmwillbeused. evaluator,sinceitdoesproduceasortedlistofgas.inamobileenvironment,as AlthoughouroriginalintentionwastoreplacethecurrentlyusedDFSalgorithm showedthatonealgorithmdoesnotmeetalltheneedsofamobilegaevalua- waythattheerthresholdwassetforthequerywindow(seesubsection3.1.2).as supportstation),ortheuserwantstoexaminealargenumberofgas,aswitch toa\focus"dfsalgorithmwouldfacilitatetheexplorationofthedatabaseusing moreresources.therefore,gaevaluatorthresholdscanbesetinmuchthesame host(e.g.,whenthemobilehostisstationaryandattachedtodockingmobility- abetterresponse-time.finally,whenevertherearefewrestrictionsonthemobile changesdynamicallyandcapturesthestatusofthemobilehost,thesearchtechniqueusedbythegaevaluatorwillalsochangetoaccommodatetheselimitations. However,asthelimitationsofthemobilehostarerelaxed,theGAevaluatorwill switchtoabfsthatcanfacilitateabroader,sweepingsearchofthegraphwith WearecurrentlyinvestigatingthisintegratedmobileGAevaluator.
19 graphicaluserinterfaces,anddatamodeling. 7.RelatedWork Mostoftheworkrelatedtoourapproachhasbeendoneintheareasofdatabase 19 highexpressivepowerofthequerylanguage[7,20].mostoftheproposedsystems Usersdonotneedtorememberattributenamesorvariablenames.Queriesare ducedwiththepurposeoffacilitatinguserinteractionwhilestillmaintainingthe quested.g+makesuseofadiagrammaticparadigmbyusingagraphwhoseedges [35]andG+[11].InQBE,thequeryismadebyllingintemplatesofrelations. adoptform,tabular,ordiagrambasedvisualparadigms.earlyexamplesofthese typesofvisualquerylanguagesthatusearelationalexternaldatamodelareqbe speciedbytypingexampletuplesexpressingtheinformationthatisbeingre- Intheareaofgraphicaluserinterfacesalargeamountofresearchhasbeenpro- correspondtothetuplesinarelation. vidingtheuserwithamoreabstractlogicalviewofthedata.ofthese,theer graphicalvisualquerysystemsthatprovidetheuserwithanerdiagramofthe model[10]isoftenusedastheexternaldatamodelinexistingvisualquerysystems. GORDAS[12],QBD[5],GRAQULA[28,33],andGQL/ER[34]areexamplesof schema.queriesinthesesystemsareformulatedbydrawingnodesandedgestobe conditionsandprojectionsarespeciedasannotationsofthenodesandedges.for ERschemadiagramwithcertainnodesandedgesreplicatedasnecessary.Selection matchedintheschemadiagram.thatis,queriesarespeciedassubgraphsofthe Semanticdatamodelsgoevenfurtherthantherelationalmodelintermsofpro- ofinterest,asimpliedhierarchicaldiagramoftheschemaisprovidedinorderto example,ingordasandqbd*,onceauserselectstheentitiesandrelationships theermodelwithoutthesupportofaggregationorquantication.picasso[19], ontheotherhand,providesanexternaluniversalrelationdatamodelthatinterfacesauniversalrelationaldatabasesystem.inpicasso,maximalobjectsare aidintheformulationofqueries.ingeneral,thedierencebetweenthesesystems cursion.further,gql/ercombinesfeaturesoftheuniversalrelationalmodeland istheirvaryingsupportinthespecicationofaggregation,quantication,andre- operators,userscandynamicallyconstructportionsofthedatabaseschemaand perusetheschemaforrelatedinformation,havingcompleteaccesstoanunderlying querysystemwhichasqbiisbasedonarichersemanticmodelthantheermodel ableand,therefore,nocharacter-typetuplevariablesarenecessary.anothervisual representedashyperedgesinahypergraphwhichcontainstextualattributelabels. semanticdatalanguage.similartoqbi,thisperusalisnotperformednavigationallybutsemantically.asopposedtoqbi,skiisdiagrambasedandsupports theselectionofaggregateandsetoperatorsaswellascomparisonoperatorsused inpredicateformulation.addingahyperedgewiththemousecreatesatuplevari- anduniversalrelationalmodelisski[21].inski,bymeansofasetofsemantics Queriesareformulatedviamouseclickswhichrevealpop-upmenusthatallowfor navigationthroughthepathsoftheunderlyingdatabaseschema.
20 userswithlittle,ifany,trainingindatabases,andskilledones.thetwoperformancemeasuresusedwerethetimeinsecondstocompleteaqueryandaccuracyof thequery.eachgroupofusersafterashorttrainingsessionofequaltimesinusing ationofqbiandqbd[24,6].inthisstudyuserswereclassiedintounskilled withrespecttounsophisticateduserswasestablishedthroughanempiricalevalu- 20Comparedtodiagrammaticvisuallanguages,theeaseandeectivenessofQBI language.usersweregiventhesequeriesindierentordersinordertominimizethe QBIandQBDweregivensixqueriesofdierentlevelsofcomplexityinnatural learningeect.ingeneral,unskilledusersdidbetterwithqbiwhereasskilledones feltmorecomfortableusingqbd,particularlyinexpressingqueriescharacterized byahighsemanticdistancevalueinvolvingpathsoflength4,ormore,andwithno cycles.thereasonwasthatskilledusersperceivedthewholepathnotasasingle acleardistinctionamongdierentoccurrencesofthesameconcept. ofthesameconcept(entityorrelationship).onthecontrary,inqbiapathcorrespondstoagaandeverygaisvisuallyrepresentedasadierenticononthe andcontrolled.ontheotherhand,therewasasignicantdierenceinaccuracy complexfunction,i.e.,ga,butasasequenceofstepsthatcanbemanuallybuilt andperformanceforquerieswithlowsemanticdistancevalueorqueriesinvolving screen.therefore,whenaqueryexpressioncontainscycles,theuserstillperceives cycles.inthepresenceofcycles,qbdusersgetmuchmoreconfusedbecausethey seemultiplecopiesofthesameform,eachcorrespondingtoadierentoccurrence palmtopcomputerandthelimitedpossibilityofusingakeyboard,maketheiconic particular,othersystemsdonotusuallyassignuniformsemanticstoicons.also, approachparticularlysuitablefortheusersofamobilesystem.ingeneral,themain asopposedtoqbi,thesesystemsadopttheextensionalbrowsingapproach(that evidentthatagreateramountofworkhasbeenperformedusingformanddiagrammaticparadigms.however,thesmallscreenspaceofthetypicalnotebookor Whencomparedtotheworkdoneinvolvingiconbasedvisualparadigms,itis dierencebetweenqbiandtheothericonicinterfacesproposedintheliterature [14,30,29]isinthewayiconsaredenedandusedforexpressingconcepts.In links. is,browsingofinstancesintheremotedatabase)astheprincipalqueryingstrategy [27,29]hencemakingthemunsuitableformobileenvironmentsthatarecharacterizedbylowcommunicationbandwidthoverexpensivewirelesscommunication needforanalternativevisualqueryparadigmformobile,pen-basedcomputersthat takesintoconsiderationtherequirementsofmobileusers,suchasexplorationof screenandnokeyboard,wasrstidentiedin[3].asopposedtoqbi,theproposed alargedatabaseschema,andthelimitationsofmobilecomputers,suchassmall datamodelwhichusestheuniversalrelationapproachatdierentlevelstocoalesce alternativeisform-basedwhereastheexternaldatamodelisamulti-levelsemantic relatedinformationandeliminatelow-levelinformationnotrelevanttotheuser.as Allvisualqueryinterfacesdiscussedabovehavebeenproposedinthecontext ofworkstationswithlargescreens,graphicscapabilitiesandpointingdevices.the statedearlier,theconceptofgasinqbiservesasimilarpurpose,coalescingrelated
21 informationofanobjectfromtheperspectiveoftheuserandrepresentingitonthe screenwithanicon. strictive,anumberofauthorshaveproposedvarioussolutions.in[18],pathex- pressionsareexaminedandformthebasisofthexsqlsystem.xsqlallowsthe specicationofpathvariablesbymeansofwhichincompletepathexpressionscan Recognizingthatquerylanguageswhichrequirefullyspeciedpathsaretoore- 21 bespecied.in[17],pathexpressionsareconsideredtobeabbreviatedqueries withinauserinterfacetoadatabasesystem.givenanambiguouspathexpression whichcouldresultinmultiplepossiblepaths,thetaskistondthosecompletions mostlikelyintendedbytheuser. PICASSO[19]providesagraphicalqueryinterfacebasedontheUniversalRelation basicrelationthattheuserhasinmind.thisrelationiscomputedthroughthe adecisionproblemconcerningwhichisthemostmeaningfulattributeistackled. WindowFunctiononthesetofattributenamesX.WithinaWindowFunction Model.Thechoiceofassigningameaningtoanattributenameisbasedonthe embeddedinattributenamesandforeverysetofattributesxthereisaunique analysisoftheschemaoftheunderlyingdatabaseandvariouskindsofdependencies. Withthisapproach,thesameattributenamecanhavedierentmeaningsifused TheideaoftheUniversalRelationModel[23,22,31]isthataccesspathsare uatingthesemanticsoftheattributes,iscommontobothqbiandtheuniversal Relationapproach.However,thequeryingstrategy,isslightlydierent.Insteadof internalschemaofthedatabase,she/hemustbeawareofthedomainofinterest. indierentcontexts;asaconsequence,eveniftheuserisnotrequiredtoknowthe thesemanticdistancefunctiontopresenttotheuserallthemeaningfulattributes beforethequeryiscomposed.moreover,theuseofasemanticmodelandstatisticalinformationonthedatabaseextensionallowsthedenitionofarichernotion ofmeaningfulnessofanattribute. Theideaofpresentingtotheuserasimpliedstructureofthedatabasebyeval- assigningameaningtoanattributeafterthequeryhasbeenspecied,qbiuses 8.Conclusions Inthispaper,wehavedescribedanicon-basedqueryprocessingfacilitycalledQBI, suitableformobileusers.thatis,qbisatisesallthreeofthecriteriaidentied intheintroductionforaneectivemobilequeryprocessingfacility: languagedoesnotinvolvepathspecicationincomposingaquery.thus,itis sizelimitationsofamobilecomputerwhilenewrequirementsarenotimposed. (1)QBIallowstheconstructionofadatabasequerywithnospecialknowledge ofhowthedatabaseisstructuredandwhereitislocated.itsiconicvisualquery equallyusefultobothunsophisticatedandexpertmobileusers. (2)Usersprimarilyinteractwiththesystemwithapointingdevice,suchasa penoramouse,andcomposeaquerybyarrangingicons.thus,itovercomesany
22 22(3)QBI'salgorithms,particularlythemetaquerytoolsandGAevaluator,are inextendingaspectsofthisworkinordertominimizetheamountofretrievedand transmitteddataoverwirelesslinks. resourceconstraintssuitableformobilequeryprocessing.further,weareinterested Asmentionedabove,wearelookingintointegratedpathcomputationsunder batterypower,andrestrictedwirelesscommunicationbandwidth. designedtoeectivelyoperateunderlimitedmemoryanddiskcapacity,limited ThisworkwassupportedinpartbyNationalScienceFoundationundergrants Acknowledgments participationinthedevelopmentoftheprototype. IRI andIRI (USA),IntegraSistemiInterattivi,andComputer andmicroimages.p.a.(italy).wealsothanks.pavaniandl.saladinifortheir References 3.AlonsoR.,E.Haber,andH.Korth.ADatabaseInterfaceforMobileComputers.Proceedings 5.AngelaccioM.,T.Catarci,andG.Santucci.QBD*:AGraphicalQueryLanguagewith 4.AlonsoR.,andH.Korth.DatabaseIssuesinNomadicComputing.ProceedingsofACM 2.AgrawalR.,S.Dar,andH.Jagadish.DirectTransitiveClosureAlgorithms:Designand 1.AbiteboulS.,andA.Bonner.ObjectsandViews.ProceedingsoftheInt'lConferenceACM- Recursion.IEEETransactionsonSoftwareEngineering,16(10): ,1990. PerformanceEvaluation.ACMTransactiononDatabaseSystems,15(3): ,1990. ofthe1992globecomworkshoponnetworkingofpersonalcommunicationapplications, SIGMODInt'lConferenceonManagementofData,pp ,May1993. Dec SIGMOD,Denver,Colorado,pp ,June BatiniC.,T.Catarci,M.F.CostabileandS.Levialdi.VisualQuerySystems.TechnicalReportNo DipartimentodiInformaticaeSistemistica,Universita'diRoma\Lgrammaticvs.anIconicQueryLanguage.(submittedforpublication),Feb Sapienza",Mar Germany,pp ,Oct icallanguage.proceedingsofthe11thint'lconferenceonentity-relationshipapproach, 8.BonoG.,andP.Ficorilli.NaturalLanguageRestatementofQueriesExpressedinaGraph- 6.BadreA.N.,T.Catarci,A.Massari,andG.Santucci.ComparativeEectivenessofaDia- 9.CatarciT.,andG.Santucci.FundamentalGraphicalPrimitivesforVisualQueryLanguages. 14.GroetteI.P.,andE.G.Nillson.SICON:anIconPresentationModuleforanE-RDatabase. 13.GallagerR.G..InformationTheoryandReliableCommunication.Wiley,NewYork, ChenP.P.TheEntityRelationshipModeltowardaUniedViewofData.ACMTransactions 12.ElmasriR.,andG.Wiederhold.GORDAS:AFormalHigh-levelQueryLanguageforthe 11.CruzI.F.,A.O.Mendelzon,andP.T.Wood.G+:RecursiveQuerieswithoutRecursion. Approach,Washington,D.C.,pp.49-72,1981. Proceedingsofthe2ndInt'lConferenceonExpertDatabaseSystems,pp ,1988. InformationSystems,3(18),pp.75-98,1993. Proceedingsofthe7thInt'lConferenceonEntityRelationshipApproach,Roma,Italy,pp ,1988. ondatabasesystems,1(1),1976. Entity-RelationshipModel.Proceedingsofthe2ndInt'lConferenceonEntity-Relationship
23 23 15.ImielinskiTandB.R.Badrinath.MobileWirelessComputing:ChallengesinDataManagement.CommunicationofACM,37(10):18-28,Oct IoannidisY.E.,R.Ramakrishnan,andL.Winger.TransitiveClosureAlgorithmsBasedon GraphTraversal.ACMTransactionsonDatabaseSystems,18(3): , IoannidisY.E.,andY.Lashkari.IncompletePathExpressionsandtheirDisambiguation. ProceedingsoftheACMSIGMODInt'lConferenceonManagementofData,Minneapolis, MI,pp ,May KiferM.,KimW.,andSagivY.QueryingObjectOrientedDatabases.Proceedingsofthe ACMSIGMODInt'lConferenceonManagementofData,pp ,May KimH.,H.Korth,andA.Silberschatz.PICASSO:AGraphicalQueryLanguage.Software PracticeandExperience,18(3): ,Mar KimW.IntroductiontoObject-OrientedDatabases.MITpress,Cambridge,MA, KingR.,andS.Melville.Ski:ASemantics-KnowledgeableInterface.Proceedingsofthe10th Int'lConferenceonVeryLargeDataBases,Singapore,pp.30-33,Aug MaierD.,D.Rozenshtein,andD.S.Warren.WindowFunctions.AdvancesInComputing Research,3: , MaierD.,andJ.D.Ullman.MaximalObjectsandtheSemanticsofUniversalRelation Databases.ACMTransactionsonDatabaseSystems,1(8):1-14, MassariA.AnIconBasedQuerySystemforRadiologicalData.Ph.D.Thesis,Dipartimento diinformaticaesistemisticauniversita'diroma"lasapienza",nov MassariA.,andP.K.Chrysanthis.VisualQueryofCompletelyEncapsulatedObjects. Proceedingsofthe5thInt'lWorkshoponResearchIssuesinDataEngineering-Distributed ObjectManagement,Taipei,Taiwan,18-25,Mar MassariA.,S.Pavani,andL.Saladini.QBI:AnIconicQuerySystemforInexpertUsers. ProceedingsoftheWorkshoponAdvancedVisualInterfaces,Bari,Italy,pp ,June MotroA.,A.D'Atri,andL.Tarantino.KIVIEW:TheDesignofanObjectOrientedBrowser. Proceedingsofthe2ndConferenceonExpertDatabaseSystems,Virginia,pp , SockutG.H.,L.M.Burns,A.Malhotra,andK.Y.Whang.GRAQULA:AGraphicalQuery LanguageforEntity-RelationshiporRelationalDatabases.ResearchReportRC16877,IBM T.J.WatsonResearchCenter,YorktownHeights,NY,Mar TonomuraY.,andS.Abe.ContentOrientedVisualInterfacesUsingVideoIconsforVisual DatabaseSystems.ProceedingsoftheIEEEWorkshoponVisualLanguages,Roma,Italy, pp.68-73, TsudaK.,M.Hirakawa,M.Tanaka,andT.Ichikawa.IconicBrowser:AnIconicRetrieval SystemforObject-Orienteddatabases.JournalofVisuallanguagesandComputing,1(1):59-76, UllmanJ.D..TheU.R.StrikesBack.ProceedingsoftheACMPrinciplesofDatabase Systems,LosAngeles,California,pp.10-22, WeissmanS..ChangingQuerybyIconstoImproveQueryingProcessingforMobileUsers. M.S.Project,UniversityofPittsburgh,May WhangK.Y.,A.Malhotra,G.H.Sockut,L.M.Burns,andK.S.Choi.Two-Dimensional SpecicationofUniversalQuanticationinaGraphicalDatabaseQueryLanguage.TransactionsonSoftwareEngineering,18(3): ,Mar ZhangZ.,andA.O.Mendelzon.AGraphicalQueryLanguageforEntityRelationship Databases.AnEntity-RelationshipApproachtoSoftwareEngineering.DavisC.,S.Jajodia, P.Ann-BengNG,andR.T.YehEds.,NorthHolland,pp , ZloofM.M.QuerybyExample.ProceedingsoftheNationalComput.Conference,pp ,1975.
Innovating for Health The Intelligent EHR
Patients and families Patients in mind Patients in your heart Innovating for Health The Intelligent EHR Kunnskapssenteret Kvalitetsindikatorer baset pa data fra medisinske kvalitetsregistre will be the
More informationAbstract. 1IntroductionandMotivation
AScalableArchitectureforAutonomous HeterogeneousDatabaseInteractions SteveMilliner,AthmanBouguettayaandMikePapazoglou fsteve,athman,mikepg@icis.qut.edu.au QueenslandUniv.ofTechnology Brisbane,QLD4001,Australia
More informationHowever,duetoboththescaleandthecomplexityoftheInternet,itisunlikelythatameasure-
Part1:AServer-BasedMeasurementInfrastructure NetworkPerformanceMeasurementandAnalysis Y.ThomasHou (ConceptPaper) AsInternettraccontinuestogrowexponentially,itisessentialforboththeusersandserviceproviders
More informationSecure Computation Martin Beck
Institute of Systems Architecture, Chair of Privacy and Data Security Secure Computation Martin Beck Dresden, 05.02.2015 Index Homomorphic Encryption The Cloud problem (overview & example) System properties
More informationSunbury Community Health Dental Service Quality and Safety Report
Sunbury Community Health Dental Service Quality and Safety Report Quality, safety and continuous improvement are at the heart of everything we do at Sunbury Community Health. We constantly work to recognise
More informationApplication-LayerAnycasting
Application-LayerAnycasting fbobby,ammar,ewz,viren,feig@cc.gatech.edu SamratBhattacharjee MostafaH.Ammar GIT{CC{96/25 EllenW.Zegura ZongmingFei VirenShah nicatewiththe\best"server,accordingtosomeperformanceorpolicycriteria,inagroup
More informationINSTALLING NOVA DATA ENCRYPTION SOFTWARE
INSTALLING NOVA DATA ENCRYPTION SOFTWARE Introduction 1. Here is what you will need to do to install the college s data encryption software on your computer, laptop, or USB storage device. The encryption
More informationSQL Simple Queries. Chapter 3.1 V3.0. Copyright @ Napier University Dr Gordon Russell
SQL Simple Queries Chapter 3.1 V3.0 Copyright @ Napier University Dr Gordon Russell Introduction SQL is the Structured Query Language It is used to interact with the DBMS SQL can Create Schemas in the
More informationBuilding Queries in Microsoft Access 2007
Building Queries in Microsoft Access 2007 Description In this class we will explore the purpose, types and uses of Queries. Learn to design a query to retrieve specific data using criteria and operators.
More informationEHR Search Engine SIBM. SJ. Darmoni, R. Lelong, C. Cabot, B. Dahamna
EHR Search Engine SIBM SJ. Darmoni, R. Lelong, C. Cabot, B. Dahamna RAVEL project, funded by French Research Agency (Technologies for Health Program) 1 Context and projects Establishment of EHR (Electronics
More informationXYZ Medica Inc. Change Management
XYZ Medica Inc. Change Management December 2006 Suggested additions to this basic report template Classify Changes as per the recommendations of ITIL then report against the separate classifications. This
More informationNew Car $16,000 5 yr. payments Car note 266.70/month. New Car $30,000 5 yr. payments Car note $500./month Car insurance $250/month Gasoline $75/week
Gasoline $75/week Car insurance $250/month Gasoline $75/week Car insurance $250/month Car insurance $250/month Gasoline $90/week Gasoline $60/week Car insurance $80/month Car insurance $250/month Car insurance
More informationStitched in a Security Fabric: Big Data Veracity NIST BGWG
Stitched in a Security Fabric: Big Data Veracity NIST BGWG Mark Underwood CEO Business Process Architect Mark.underwood@kryptonbrothers.com DHS/DOD Cybersecurity R&D CEP AI JIT Learning Business Intelligence
More informationOntology-Based Semantic Modeling of Safety Management Knowledge
2254 Ontology-Based Semantic Modeling of Safety Management Knowledge Sijie Zhang 1, Frank Boukamp 2 and Jochen Teizer 3 1 Ph.D. Candidate, School of Civil and Environmental Engineering, Georgia Institute
More informationLocalErrorRecoveryinSRM: ComparisonofTwoApproaches. Ching-GungLiu,DeborahEstrin,ScottShenkerandLixiaZhang
LocalErrorRecoveryinSRM: ComparisonofTwoApproaches Ching-GungLiu,DeborahEstrin,ScottShenkerandLixiaZhang AbstractSRMisaframeworkforreliablemulticastdelivery.Inordertomaximizethecollaborationamongthe groupmembersinerrorrecovery,bothretransmissionrequestsandrepliesaremulticasttotheentiregroup.while
More informationBagged Ensemble Classifiers for Sentiment Classification of Movie Reviews
www.ijecs.in International Journal Of Engineering And Computer Science ISSN:2319-7242 Volume 3 Issue 2 February, 2014 Page No. 3951-3961 Bagged Ensemble Classifiers for Sentiment Classification of Movie
More informationA neo4j powered social networking and Question & Answer application to enhance scientific communication. René Pickhardt, Heinrich Hartmann
A neo4j powered social networking and Question & Answer application to enhance scientific communication. René Pickhardt, Heinrich Hartmann related-work.net Roadmap Introduction Data structures for Q &
More informationTopics. Introduction to Database Management System. What Is a DBMS? DBMS Types
Introduction to Database Management System Linda Wu (CMPT 354 2004-2) Topics What is DBMS DBMS types Files system vs. DBMS Advantages of DBMS Data model Levels of abstraction Transaction management DBMS
More informationXYZ Medica Inc. Incident Management
XYZ Medica Inc. Incident Management December 2006 Suggested additions to this basic report template Classify Incidents as per the recommendations of ITIL then report against the separate classifications.
More informationNames & Addresses. Names & Addresses. Names vs. Addresses. Identity. Names vs. Addresses. CS 194: Distributed Systems: Naming
Names & Addresses CS 9: Distributed Systems: Naming Computer Science Division Department of Electrical Engineering and Computer Sciences University of California, Berkeley Berkeley, CA 970-77 What is a?
More informationURI and UUID. Identifying things on the Web.
URI and UUID Identifying things on the Web. Overview > Uniform Resource Identifiers (URIs) > URIStreamOpener > Universally Unique Identifiers (UUIDs) Uniform Resource Identifiers > Uniform Resource Identifiers
More informationFINPP07 Support the ongoing client relationship
Overview This unit is for those who provide paraplanning services to financial planners in the delivery of bespoke solutions to clients as part of the financial advice process. You must be able to support
More informationAn Approach for Knowledge-Based IT Management of Air Traffic Control Systems
An Approach for Knowledge-Based IT Management of Air Traffic Control Systems Fabian Meyer, Reinhold Kroeger RheinMain University of Applied Sciences D-65195 Wiesbaden, Germany {firstname.lastname}@hs-rm.de
More informationClinical Mapping (CMAP) Draft for Public Comment
Integrating the Healthcare Enterprise 5 IHE Patient Care Coordination Technical Framework Supplement 10 Clinical Mapping (CMAP) 15 Draft for Public Comment 20 Date: June 1, 2015 Author: PCC Technical Committee
More informationMath 58. Rumbos Fall 2008 1. Solutions to Assignment #3
Math 58. Rumbos Fall 2008 1 Solutions to Assignment #3 1. Use randomization to test the null hypothesis that there is no difference between calcium supplementation and a placebo for the experimental data
More informationNext Generation Requirements Engineering
Next Generation Requirements Engineering Silvia Mazzini, Intecs SPA (silvia.mazzini@intecs.it ) 1 The NextGenRE Project Next Generation Requirements Engineering (NextGenRE) is a study funded by ESA/ESTEC
More informationThe Interface Concept
Multiple inheritance Interfaces Four often used Java interfaces Iterator Cloneable Serializable Comparable The Interface Concept OOP: The Interface Concept 1 Multiple Inheritance, Example Person name()
More informationSemantically Steered Clinical Decision Support Systems
Semantically Steered Clinical Decision Support Systems By Eider Sanchez Herrero Department of Computer Science and Artificial Intelligence University of the Basque Country Advisors Prof. Manuel Graña Romay
More informationIP Phone Service Administration and Subscription
CHAPTER 6 IP Phone Service Administration and Subscription Cisco CallManager administrators maintain the list of services to which users can subscribe. These sections provide details about administering
More informationContemporary Self Contained Office Mews To Let - Highgate Village, London N6
Pound House 62a Highgate High Street, Highgate, LONDON, N6 5HX Area Net Internal Area: 136 sq.m. (1,466 sq.ft.) Rent Rent on application Property Description The property comprises the ground and upper
More informationRecipes for Building Materially Diverse, Functionally Transparent and Expressive Electronics
RecipesforBuildingMateriallyDiverse,Functionally TransparentandExpressiveElectronics HannahPerner Wilson ThesisProposalfortheDegreeofMasterofScienceinthePrograminMediaArtsand SciencesattheMassachusettsInstituteofTechnology,Fall2010
More information11 November 2015. www.isbe.tue.nl. www.isbe.tue.nl
UML Class Diagrams 11 November 2015 UML Class Diagrams The class diagram provides a static structure of all the classes that exist within the system. Classes are arranged in hierarchies sharing common
More informationBeyond Market Research
Beyond Market Research A New Look At Defining The Insights Sector Harry Henry VP & Lead Analyst IIeX 2014 June, 2014 Atlanta, GA A d v a n c i n g t h e B u s i n e s s o f I n f o Ar md va at ino cn i
More informationA pixlogic White Paper 4984 El Camino Real Suite 205 Los Altos, CA 94022 T. 650-967-4067 info@pixlogic.com www.pixlogic.com
A pixlogic White Paper 4984 El Camino Real Suite 205 Los Altos, CA 94022 T. 650-967-4067 info@pixlogic.com www.pixlogic.com Intelligent License Plate Recognition for Security Applications Dec-2011 Contents
More informationCALIFORNIA DEPARTMENT OF INSURANCE ENABLING ENTERPRISE MANAGEMENT via the ENTERPRISE INFORMATION PORTAL
CALIFORNIA DEPARTMENT OF INSURANCE ENABLING ENTERPRISE MANAGEMENT via the ENTERPRISE INFORMATION PORTAL This nomination is being submitted under the award category of Digital Government Government to Business
More information5 Point Choice ( 五 分 選 擇 題 ): Allow a single rating of between 1 and 5 for the question at hand. Date ( 日 期 ): Enter a date Eg: What is your birthdate
5 Point Choice ( 五 分 選 擇 題 ): Allow a single rating of between 1 and 5 for the question at hand. Date ( 日 期 ): Enter a date Eg: What is your birthdate Gender ( 性 別 ): Offers participants a pre-defined
More informationI/O-Efficient Data Structures for Colored Range and Prefix Reporting
I/O-Efficient Data Structures for Colored Range and Prefix Reporting Kasper Green Larsen, MADALGO, Aarhus University Rasmus Pagh, IT University of Copenhagen Presenter: Yakov Nekrich 1 Motivating example
More informationSouth Australia State Tax Review. February 2015
South Australia State Tax Review February 2015 On Wednesday 11 February 2015, the South Australian government released a discussion paper on reform to the South Australian taxation system. The discussion
More informationRelational Database Basics Review
Relational Database Basics Review IT 4153 Advanced Database J.G. Zheng Spring 2012 Overview Database approach Database system Relational model Database development 2 File Processing Approaches Based on
More informationBigData in Real-time. Impala Introduction. TCloud Computing 天 云 趋 势 孙 振 南 zhennan_sun@tcloudcomputing.com. 2012/12/13 Beijing Apache Asia Road Show
BigData in Real-time Impala Introduction TCloud Computing 天 云 趋 势 孙 振 南 zhennan_sun@tcloudcomputing.com 2012/12/13 Beijing Apache Asia Road Show Background (Disclaimer) Impala is NOT an Apache Software
More informationCS 564: DATABASE MANAGEMENT SYSTEMS
Fall 2013 CS 564: DATABASE MANAGEMENT SYSTEMS 9/4/13 CS 564: Database Management Systems, Jignesh M. Patel 1 Teaching Staff Instructor: Jignesh Patel, jignesh@cs.wisc.edu Office Hours: Mon, Wed 1:30-2:30
More informationAn Intuitive Graphical Query Interface for Protégé Knowledge Bases
An Intuitive Graphical Query Interface for Protégé Knowledge Bases Landon Todd Detwiler, Cornelius Rosse, Linda Shapiro Structural Informatics Group University of Washington Emily Query interface for Protégé
More informationVeterans' Organizations
General Revenue Fund Group 743501 American Ex-Prisoners of War Appropriation Appropriation $27,533 $27,533 $28,910 $28,910 $28,910 $28,910 by Am. Sub. H.B. 215 of the 122nd G.A.) 746501 Army and Navy Union,
More informationMotor Vehicle Injuries
Motor Vehicle Injuries Prenatal Counseling about Seat Belt Use during Pregnancy and Injuries from Car Crashes during Pregnancy Background The CDC has identified prevention of motor vehicle injuries as
More informationPERSONAL PROTECTIVE EQUIPMENT
PERSONAL PROTECTIVE EQUIPMENT Background The Safety, Health and Welfare at Work Act 2005 applies to employers, employees in all employments and to the self employed. The Act contains provisions for improving
More informationCambridge Medical & Rehabilitation Center Tod N. Lambert
Cambridge Medical & Rehabilitation Center Tod N. Lambert Group Chief Executive Officer Cambridge Medical & Rehabilitation Center Post Acute Continuum of Care Transport Prevention & Early Detection Family
More informationHealth History Form: Beverly Hosford, Muscle Activation Techniques Specialist
Health History Form: Beverly Hosford, Muscle Activation Techniques Specialist Name: Phone: Address: E-mail: Birthdate: Age: EMERGENCY CONTACT Name: Phone: Are you seeing any medical practitioners such
More informationConfiguring Sites and Understanding AD replication. Dante Villarroel Saavedra
Configuring Sites and Understanding AD replication Dante Villarroel Saavedra Agenda Introduction Understanding Sites Sites planning Active Directory Partitions Global Catalog Active Directory Replication
More informationVISUALIZATION OF GEOMETRICAL AND NON-GEOMETRICAL DATA
VISUALIZATION OF GEOMETRICAL AND NON-GEOMETRICAL DATA Maria Beatriz Carmo 1, João Duarte Cunha 2, Ana Paula Cláudio 1 (*) 1 FCUL-DI, Bloco C5, Piso 1, Campo Grande 1700 Lisboa, Portugal e-mail: bc@di.fc.ul.pt,
More informationMining Signatures in Healthcare Data Based on Event Sequences and its Applications
Mining Signatures in Healthcare Data Based on Event Sequences and its Applications Siddhanth Gokarapu 1, J. Laxmi Narayana 2 1 Student, Computer Science & Engineering-Department, JNTU Hyderabad India 1
More informationFixed Assets Process Guide
Fixed Assets Process Guide SAP (& SUN) This Guide represents current processes about how the Fixed Asset Process will be operated. This guide has been produced in consultation with FABS, Internal Audit
More informationTechnical Writing Examples Plus A Few Tips
Technical Writing Examples Plus A Few Tips What is wrong? How to rewrite? Before The volume of information has been rapidly increasing in the past few decades. While computer technology has played a significant
More informationBig Data Processing Can We Control the Value of Lost Data?
Expert Panel ALLDATA/MMEDIA/KESA Wednesday, April 22, 5:30 pm Barcelona Big Data Processing Can We Control the Value of Lost Data? Moderator Philip Davies, Bournemouth University, UK Panelists Venkat N.
More informationCHAPTER-12. Analytical Characterization : Analysis of Attribute Relevance
CHAPTER-12 Analytical Characterization : Analysis of Attribute Relevance 12.1 Introduction 12.2 Methods of Attribute Relevance Analysis 12.3 Review Questions 12.4 References 12. Analytical Characterization
More informationSoftware Engineering. System Models. Based on Software Engineering, 7 th Edition by Ian Sommerville
Software Engineering System Models Based on Software Engineering, 7 th Edition by Ian Sommerville Objectives To explain why the context of a system should be modeled as part of the RE process To describe
More informationHuman Interactions in Dynamic Environments through Mobile Web Services
Human Interactions in Dynamic Environments through Mobile Web Services Daniel Schall, Robert Gombotz, Christoph Dorn, Schahram Dustdar Distributed Systems Group Institute of Information Systems Vienna
More informationExpert Systems : AI Course Lecture 35 36, notes, slides www.myreaders.info/, RC Chakraborty, e-mail rcchak@gmail.
Expert Systems : AI Course Lecture 35 36, notes, slides www.myreaders.info/, RC Chakraborty, e-mail rcchak@gmail.com, June 01, 2010 www.myreaders.info/html/artificial_intelligence.html www.myreaders.info
More informationFlexible and modular visualisation and data discovery tools for environmental information
21st International Congress on Modelling and Simulation, Gold Coast, Australia, 29 Nov to 4 Dec 2015 www.mssanz.org.au/modsim2015 Flexible and modular visualisation and data discovery tools for environmental
More informationEnergy: 4.C.1 Introduction to Energy
Partnerships Implementing Engineering Education Worcester Polytechnic Institute Worcester Public Schools Supported by: National Science Foundation Energy: 4.C.1 Introduction to Energy Grade Level 4 Sessions
More informationACL Based Dynamic Network Reachability in Cross Domain
South Asian Journal of Engineering and Technology Vol.2, No.15 (2016) 68 72 ISSN No: 2454-9614 ACL Based Dynamic Network Reachability in Cross Domain P. Nandhini a, K. Sankar a* a) Department Of Computer
More informationIMPROVING BUSINESS PROCESS MODELING USING RECOMMENDATION METHOD
Journal homepage: www.mjret.in ISSN:2348-6953 IMPROVING BUSINESS PROCESS MODELING USING RECOMMENDATION METHOD Deepak Ramchandara Lad 1, Soumitra S. Das 2 Computer Dept. 12 Dr. D. Y. Patil School of Engineering,(Affiliated
More informationHPI in-memory-based database system in Task 2b of BioASQ
CLEF 2014 Conference and Labs of the Evaluation Forum BioASQ workshop HPI in-memory-based database system in Task 2b of BioASQ Mariana Neves September 16th, 2014 Outline 2 Overview of participation Architecture
More informationHow To Trade In Koromand
LIQUIDITY PROVIDER FEE LIST Table of Contents FEE LIST MEMBER S PROPRIETARY TRADING - LIQUIDITY PROVIDER EQUITY DERIVATIVES DANISH DERIVATIVES NORWEGIAN DERIVATIVES SWEDISH DERIVATIVES FEE LIST CUSTOMER
More informationPrediction of Heart Disease Using Naïve Bayes Algorithm
Prediction of Heart Disease Using Naïve Bayes Algorithm R.Karthiyayini 1, S.Chithaara 2 Assistant Professor, Department of computer Applications, Anna University, BIT campus, Tiruchirapalli, Tamilnadu,
More informationNumerical Research on Distributed Genetic Algorithm with Redundant
Numerical Research on Distributed Genetic Algorithm with Redundant Binary Number 1 Sayori Seto, 2 Akinori Kanasugi 1,2 Graduate School of Engineering, Tokyo Denki University, Japan 10kme41@ms.dendai.ac.jp,
More information02-201: Programming for Scientists
1. Course Information 1.1 Course description 02-201: Programming for Scientists Carl Kingsford Fall 2015 Provides a practical introduction to programming for students with little or no prior programming
More informationBo Holland is the Founder & CEO of AllClear ID with deep expertise in Identity theft prevention, security, and technology
AllClear ID (www.allclearid.com) is an identity protection company based in Austin, Texas, and is powered by Debix, a leader in identity protection technology. Our leading technology and customer service
More informationIntroduction to Databases
Introduction to Databases IT University of Copenhagen January 7, 2005 This exam consists of 6 problems with a total of 16 questions. The weight of each problem is stated. You have 4 hours to answer all
More informationA = P (1 + r / n) n t
Finance Formulas for College Algebra (LCU - Fall 2013) ---------------------------------------------------------------------------------------------------------------------------------- Formula 1: Amount
More informationLecture 13. Understanding Probability and Long-Term Expectations
Lecture 13 Understanding Probability and Long-Term Expectations Thinking Challenge What s the probability of getting a head on the toss of a single fair coin? Use a scale from 0 (no way) to 1 (sure thing).
More informationFutureWise at a glance. Macquarie Life
FutureWise at a glance Macquarie Life Macquarie Life FutureWise offers a comprehensive and competitively priced suite of products that enables you to choose the right combination of cover types and benefits
More informationetapestry Data Enrichment Additional Import Options
etapestry Data Enrichment Additional Import Options etapestry now has the following Additional Import Options: Residential Phone Append Business Phone Append Canadian Phone Append County Append Deceased
More informationMaster of Science Business Information Systems. Agile Processes Combining Business Processes and Business Rules
Master of Science Business Information Systems Agile Processes Combining Business Processes and Business Rules Knowledge and Processes knowledge about processes: workflow roles process logic used at design
More informationYale University. Medicaid Calculator. Documentation
Yale University Medicaid Calculator Documentation To Accompany Medicaid as an Investment in Children: What is the Long-Term Impact on Tax Receipts?" by David Brown, Amanda Kowalski, Ithai Lurie 11-21-
More informationHorizontal IoT Application Development using Semantic Web Technologies
Horizontal IoT Application Development using Semantic Web Technologies Soumya Kanti Datta Research Engineer Communication Systems Department Email: Soumya-Kanti.Datta@eurecom.fr Roadmap Introduction Challenges
More informationOverview and Legal Context
Impact of ACOs on Physician/Provider Membership Decisions 0 Overview and Legal Context Michael R. Callahan Katten Muchin Rosenman LLP Vice Chair, Medical Staff Credentialing and Peer Review Practice Group
More informationMuhadith: A Cloud based Distributed Expert System for Classification of Ahadith
2012 10th International Conference on Frontiers of Information Technology Muhadith: A Cloud based Distributed Expert System for Classification of Ahadith Kashif Bilal North Dakota State University, Fargo,
More informationGerald Hiebel 1, Øyvind Eide 2, Mark Fichtner 3, Klaus Hanke 1, Georg Hohmann 4, Dominik Lukas 5, Siegfried Krause 4
OGC GeoSparql and CIDOC CRM Gerald Hiebel 1, Øyvind Eide 2, Mark Fichtner 3, Klaus Hanke 1, Georg Hohmann 4, Dominik Lukas 5, Siegfried Krause 4 1 Surveying and Geoinformation Unit, University of Innsbruck
More informationAPPENDIX A: OBJECTIVES AND MEASURES FOR 2015 THROUGH 2017 (MODIFIED STAGE 2) EP Objectives and Measures
APPENDIX A: OBJECTIVES AND MEASURES FOR 2015 THROUGH (MODIFIED STAGE 2) Objectives for Measures for Providers in EP Objectives and Measures Objective 1: Protect Patient Health Information Objective 2:
More informationInformation Systems (Informationssysteme)
Information Systems (Informationssysteme) Jens Teubner, TU Dortmund jensteubner@cstu-dortmundde Summer 2015 c Jens Teubner Information Systems Summer 2015 1 Part III A Very Brief Introduction to SQL c
More informationMODULE C ONLINE SOFTWARE USER MANUAL FOR UPDATING MASTER DATA
MODULE C ONLINE SOFTWARE USER MANUAL FOR UPDATING MASTER DATA USER-MANUAL Version 1.0 Integrated Disease Surveillance Project (IDSP) National Informatics Centre Department of Information Technology Ministry
More informationLectures 2 & 3: Introduction to Modeling & UML. Getting started
Lectures 2 & 3: Introduction to Modeling & UML Why Build Models? What types of Models to build Intro to UML Class Diagrams Relationship between UML and program code Uses of UML 202 Steve Easterbrook. This
More informationMaximizing Social Security Retirement Benefits for Married Baby Boomers
Volume 3, Issue 3 Maximizing Social Security Retirement Benefits for Married Baby Boomers On the final day of the 2012 NAELA Annual Conference in Seattle in April, Attorney David A. Cechanowicz, JD, MSFS
More information800-424-1716. A global leader in: Garner Environmental Services: www.garner-es.com. Garner Decontamination Solutions: www.garner-ds.
800-424-1716 A global leader in: Emergency Response Disaster Response Biological Decontamination Industrial Services Training & Rescue Health & Safety Garner Environmental Services: www.garner-es.com Garner
More informationWHY IS BREATHING SO IMPORTANT?
You need to breathe to stay alive. But how clean is the air you breathe? If you breathe dirty air, you are more likely to develop health problems and become ill. Plants and animals need clean air too.
More informationLecture 3 Topics on Requirements Engineering
Lecture 3 Topics on Requirements Engineering Some material taken from the Tropos project at U of T Copyright Yijun Yu, 2005 Course information Let s vote Course Project/Final Exam 50-50 or 60-40? Midterm/Final
More informationComparison of Classification Techniques for Heart Health Analysis System
International Journal of Computer Sciences and Engineering Open Access Review Paper Volume-04, Issue-02 E-ISSN: 2347-2693 Comparison of Classification Techniques for Heart Health Analysis System Karthika
More information2. Conceptual Modeling using the Entity-Relationship Model
ECS-165A WQ 11 15 Contents 2. Conceptual Modeling using the Entity-Relationship Model Basic concepts: entities and entity types, attributes and keys, relationships and relationship types Entity-Relationship
More informationCourse Outline. Configuring, Managing & Maintaining Windows 2008 Server. Course Description: Pre-requisites:
Configuring, Managing & Maintaining Windows 2008 Server Course Description: This five-day instructor-led course combines five days worth of instructor-led training content from the Network Infrastructure
More informationSouth Carolina Department of Social Services CUSTODIAL PARENT S APPLICATION FOR CHILD SUPPORT SERVICES. Child Support Services
South Carolina Department of Social Services CUSTODIAL PARENT S APPLICATION FOR CHILD SUPPORT SERVICES The disclosure of your Social Security Number is mandatory, in accordance with section 466(a)(13)
More informationIntroduction to HDFS. Prasanth Kothuri, CERN
Prasanth Kothuri, CERN 2 What s HDFS HDFS is a distributed file system that is fault tolerant, scalable and extremely easy to expand. HDFS is the primary distributed storage for Hadoop applications. Hadoop
More informationData Cleansing for Remote Battery System Monitoring
Data Cleansing for Remote Battery System Monitoring Gregory W. Ratcliff Randall Wald Taghi M. Khoshgoftaar Director, Life Cycle Management Senior Research Associate Director, Data Mining and Emerson Network
More informationLanier Chiropractic and Rehabilitation Information 4530 Nelson Brogdon Blvd., Suite B, Sugar Hill, GA 30024 770-271-8949
Lanier Chiropractic and Rehabilitation Information 4530 Nelson Brogdon Blvd., Suite B, Sugar Hill, GA 30024 770-271-8949 Thank you for choosing Lanier Chiropractic and Rehabilitation! It is our desire
More informationGuidelines on Local Property Tax Relief For Disabled / Incapacitated Individuals
Guidelines on Local Property Tax Relief For Disabled / Incapacitated Individuals May 2014 Table of Contents 1. Introduction... 2 2. Summary of new rules... 3 2.1 Reduction in the chargeable value of a
More informationThe Burden of Out-of-Pocket Costs for Canadians with Diabetes
The Burden of Out-of-Pocket Costs for Canadians with Diabetes The Issue: Government coverage of diabetes medications, devices and supplies varies across jurisdictions, leaving some costs for these supports
More informationA Hybrid Approach To Web Usage Mining
A Hybrid Approach To Web Usage Mining Authors: Søren E. Jespersen Jesper Thorhauge Torben Bach Pedersen Technical Report 02-5002 Department of Computer Science Aalborg University Created on July 17, 2002
More informationHow to Search in Banner
Please note: This document assumes that you already understand the basics of how to login and use Banner. In Banner, we enter the general information about each person/student/entity only one time. For
More informationElsevier ClinicalKey TM FAQs
Elsevier ClinicalKey TM FAQs Table of Contents What is ClinicalKey? Where can I access ClinicalKey? What medical specialties are covered in ClinicalKey? What information is available through ClinicalKey?
More informationDistributed Database Design (Chapter 5)
Distributed Database Design (Chapter 5) Top-Down Approach: The database system is being designed from scratch. Issues: fragmentation & allocation Bottom-up Approach: Integration of existing databases (Chapter
More information