【MOOC】《BigDataAnalysisBeijingInstituteofTechnology》(北京理工大學(xué))期末考試慕課答案

有些題目順序不一致，下載后按鍵盤ctrl+F進(jìn)行搜索BigDataAnalysisExamBigDataAnalysisExam1.單選題：Databaseconnectionprogramminginterfacessuchas()cansupportSQLaccessbyapplicationstothedatabase,buttheycannotprovidecomplexfunctionssuchastransactionmanagement,concurrentscheduling,buffermanagement,heterogeneousdatabaseconversionandinheritanceinadistributedcomputingenvironment.Thisintroducesthe().Itisalayerofsoftwarethatprovidesdataexchangefunctionsontopofthedatabase.Whenthesystemisextendedandneedtoaccesscross-platformheterogeneousdatabases,OScouldbeUNIX,LinuxorWindows,formscouldbemails,XMLdocuments,EJBcomponents,Webservices,images,audio/videofilesorForotherunstructureddata,Andthetechnologyofthebigdataapplicationlayerisalsodiversifiedandvariousstandards.Thedesignofthe()needstobecompatiblewithvariousstandardtechnologiesandproducts,whichintroducesthe().

選項(xiàng)：

A、ODBCandJDBC;DALdataaccesslayer;Unifieddataaccessinterface;Unifieddataaccessinterface;

B、ODBCandJDBC;DALdataaccesslayer;DALdataaccesslayer;Unifieddataaccessinterface;

C、DALdataaccesslayer;ODBCandJDBC;DALdataaccesslayer;Unifieddataaccessinterface;

D、ODBCandJDBC;DALdataaccesslayer;Unifieddataaccessinterface;DALdataaccesslayer;

答案：【ODBCandJDBC;DALdataaccesslayer;DALdataaccesslayer;Unifieddataaccessinterface;】2.單選題：NoSQLdatabasedeliverhighperformanceonaggregationquerieslikeSUM,COUNT,AVG,MINetc.

選項(xiàng)：

A、Key-valuedatabase

B、Columnbaseddatabase

C、Document-OrientedDB

D、Graphbaseddatabase

答案：【Columnbaseddatabase】3.單選題：AccordingtotheCAPtheorem,adistributedsystemcanonlymeettwooftheCAPsinthemid-rundatareadandwriteoperations,whichpropertiescanMongoDBsatisfy?

選項(xiàng)：

A、consistency(C)andavailability(A)

B、consistency(C)andpartitiontolerance(P)

C、availability(A)andpartitiontolerance(P)

D、Noneofthem

答案：【consistency(C)andpartitiontolerance(P)】4.單選題：TherightorderofreadingdatainHDFS.()a)DistributedFilesystemmakesanRPCcalltothenamenodetodeterminelocationofdatanodeswherefilesarestoredinformofblocks.Foreachblock,thenamenodereturnsaddressofdatanodes(metadataofblocksanddatanodes)thathaveacopyofblock.Datanodesaresortedaccordingtoproximity(dependingofnetworktopologyinformation).?b)Theclientopensthefilebycallingopen()methodonDistributedFilesystem.c)Theclientthencallsread()onthestream.DFSInputStream,whichhasstoredthedatanodeaddressesforthefirstfewblocksinthefile,thenconnectstothefirst(closest)datanodeforthefirstblockinthefile.d)TheDistributedFilesystemreturnsanFSDataInputStream(aninputstreamthatsupportsfileseeks)totheclientforittoreaddatafrom.FSDataInputStreaminturnwrapsaDFSInputStream,whichmanagesthedatanodeandnamenodeI/O.e)Dataisstreamedfromthedatanodebacktotheclient(intheformofpackets)andread()isrepeatedlycalledonthestreambyclient.f)Whentheclienthasfinishedreading,itcallsclose()ontheFSDataInputStreamWhentheendoftheblockisreached,DFSInputStreamwillclosetheconnectiontothedatanode,thenfindthebestdatanodeforthenextblock.

選項(xiàng)：

A、abdcegf

B、badcegf

C、badcefg

D、bacdegf

答案：【badcegf】5.單選題：Whydoweneeddatamodeling,becauseitcansupport()

選項(xiàng)：

A、Datastoragestructuredesign

B、Databasedesign

C、Calculationmodel

D、Applicationdesign

答案：【Applicationdesign】6.單選題：Datamodelingcouldincludedefining()1)Metadata2)Datastructure3)Attributes4)Valuerange5)Associationrelationship6)Consistency7)Timeliness

選項(xiàng)：

A、12345

B、1234567

C、134567

D、123567

答案：【1234567】7.單選題：5Whatattributessubsetselectionmethodshowedinthediagram?

選項(xiàng)：

A、ForwardStepwiseAttributessubsetselection

B、BackwardStepwiseAttributessubsetselection

C、Combineforwardselectionandbackwarddeletion

D、Decisiontree(decisiontree)induction

答案：【BackwardStepwiseAttributessubsetselection】8.單選題：WhichoneisNOTTensorFlowprovideshardwareaccelerators?

選項(xiàng)：

A、CPUs

B、GPUs

C、NPUs

D、TPUs

答案：【NPUs】9.單選題：WhichofthefollowingisNOTcorrectdescriptionaboutTensorFlow?

選項(xiàng)：

A、AdoptedKerasforhigh-levelAPIs;

B、EagerExecutionwithGraphs;

C、NodesandtensorsinTensorFlowarePythonobjects,andTensorFlowapplicationsarethemselvesC++applications;

D、ItusesPythontoprovideaconvenientfront-endAPIforbuildingapplicationswiththeframework,whileexecutingthoseapplicationsinhigh-performanceC++;

答案：【NodesandtensorsinTensorFlowarePythonobjects,andTensorFlowapplicationsarethemselvesC++applications;】10.單選題：TheMachineLearningPipelineinSparkMLlibis()

選項(xiàng)：

A、1.Load/CleanData,2.Transformer,3.Estimatorand4.Evaluator

B、1.Load/cleandata,2.Featureextraction,3.Modeltrainingand4.Modelevaluation

C、1.Load/cleandata,2.Featureextraction,3.Estimatorand4.Modelevaluation

D、1.Load/CleanData,2.Transformer,3.Modeltrainingand4.Evaluator

答案：【1.Load/CleanData,2.Transformer,3.Estimatorand4.Evaluator】11.單選題：WhichofthefollowingisNOTpartofDataprocessingsystem?

選項(xiàng)：

A、ComputingAlgorithm

B、ComputingModel.

C、ComputingPlatform&Engine.

D、DataCollectionandModeling.

答案：【DataCollectionandModeling.】12.單選題：Superstepexecutionprocessis()1)Sendmessagestoothernodescausingthemactive;2)Modifynodeandarcproperties;3)Removetheexistingorcreatingnewarcs;4)Receivemessagefrominbox;5)Haltselfuntilnewmessagereceived;

選項(xiàng)：

A、42513

B、12345

C、42135

D、42351

答案：【42513】13.單選題：()uses()todividetheamountofresources(CPU,memory,etc.)onthisnode.ATaskhasachancetorunafteritgetsa(),andtheroleofthe()istoallocateidle()oneach()totheTask.()

選項(xiàng)：

A、JobTracker,slot,slot,Hadoopscheduler,slots,TaskTracker;

B、TaskTracker,slot,slot,Hadoopscheduler,slots,TaskTracker;

C、TaskTracker,slot,slot,Taskscheduler,slots,TaskTracker;

D、TaskTracker,slot,slot,Hadoopscheduler,task,TaskTracker;

答案：【TaskTracker,slot,slot,Hadoopscheduler,slots,TaskTracker;】14.單選題：HANAimprovedthedataanalysisperformanceindatawarehouse,Notbecause()

選項(xiàng)：

A、Iteliminatesunnecessarycomplexityandlatency

B、Acceleratethroughsimplification

C、Leveragingthepowerofin-memorycomputingallowsHANAtobringOLTP,transactionprocessing,andOLAP,dataanalytics,backtogetherinonedatabase.

D、Specializeddatawarehousesforreportingandanalyticsrequiredthemoving,transformationandpre-processingoftransactionaldata,whichintroducesahugecomplexity:sometimesanenterprisemayholdthreedifferentcopiesofthesamedata

答案：【Specializeddatawarehousesforreportingandanalyticsrequiredthemoving,transformationandpre-processingoftransactionaldata,whichintroducesahugecomplexity:sometimesanenterprisemayholdthreedifferentcopiesofthesamedata】15.單選題：StormisaNativeStreamProcessingSystem,thatis,theprocessingofstreamdataisbasedoneachpieceofdata,anditsparallelcalculationisimplementedbasedonadirectedtopologygraph.Topologycomposedofdatasource-()andprocessingunit-().TopologyDefinesthe()ofparallelcomputing,thatis,designsthecalculationstepsandprocessesfromtheperspectiveoffunctionandarchitecture.

選項(xiàng)：

A、Bolt,Spout,physicalmodel

B、Spout,Bolt,physicalmodel

C、Bolt,Spout,logicalmodel

D、Spout,Bolt,logicalmodel

答案：【Spout,Bolt,logicalmodel】16.單選題：Onceyou’vetrainedandsavedyourmodel,youcanexecuteitdirectlyinyourapplicationorserveitusingoneofthedeploymentlibraries,whichoneisNOTsuitableforthistask?

選項(xiàng)：

A、TensorFlowServing

B、TensorFlowLite

C、TensorFlow.js

D、TensorFlowHub

答案：【TensorFlowHub】17.單選題：MLlibmaincomponentsdonotinclude()

選項(xiàng)：

A、Featurization

B、Transformation

C、Pipeline

D、Utilities.

答案：【Transformation】18.單選題：MonteCarlomethodsbelongto()algorithmcategory.

選項(xiàng)：

A、Supervisedlearning

B、Unsupervisedlearning

C、Semi-supervisedlearning

D、Reinforcementlearning

答案：【Reinforcementlearning】19.單選題：InOLTP,theuseraccessesacentraldatabase,andifthe（）systemstructureisadopted,itismuchmoreefficient.Fromtheperspectiveof()architecture,itcanintegratemanyCPUsinonephysicalserver,sothatthesystemhashighertransactionprocessingcapabilities.Sincetheremotememoryaccessdelayismuchlongerthanlocalmemoryaccess,itisnecessarytominimizethedatainteractionbetweendifferentCPUmodules.Obviously,the()architectureismoresuitablefortheOLTPtransactionprocessingenvironment.Inthedatawarehouseenvironment,duetoalargenumberofcomplexdataprocessing,whichwillinevitablyleadtoalargenumberofdatainteractions,whichwillgreatlyreducetheutilizationoftheCPU,then()structurecouldbeasolution.

選項(xiàng)：

A、SMP,NUMA,NUMA,MPP

B、MPP,NUMA,NUMA,SMP

C、SMP,SMP,NUMA,MPP

D、SMP,NUMA,MPP,MPP

答案：【SMP,NUMA,NUMA,MPP】20.單選題：Inthefollowing,whichoneissharednothingarchitecture.()

選項(xiàng)：

A、SMP

B、NUMA

C、MPP

D、Noneofthem

答案：【MPP】21.單選題：Inthefollowing,whichoneissharedeverythingarchitecture.()

選項(xiàng)：

A、SMP

B、NUMA

C、MPP

D、Noneofthem

答案：【SMP】22.單選題：InGraphcomputingthemostbasicunitisanode,whichcontains()

選項(xiàng)：

A、Node’sproperties,outwardarcsanditsproperties,alogicalinboxtoreceiveallmessagessenttoit.

B、Node’sproperties,outwardarcsanditsproperties,thenodeidthattheoutwardarcpointsto

C、Node’sproperties,outwardarcsanditsproperties,thenodeidthattheoutwardarcpointsto,alogicalinboxtoreceiveallmessagessenttoit.

D、outwardarcsanditsproperties,thenodeidthattheoutwardarcpointsto,alogicalinboxtoreceiveallmessagessenttoit.

答案：【Node’sproperties,outwardarcsanditsproperties,thenodeidthattheoutwardarcpointsto,alogicalinboxtoreceiveallmessagessenttoit.】23.單選題：InmemoryDatabaseHANAcanprocessdatafast,mainlybecause()

選項(xiàng)：

A、HANAappliedmulti-corearchitecture

B、HANAdeployedonthehigh-performanceserver

C、HANAdesignedfastindexingmechanism

D、HANAstoredthemostoftenuseddatainmemory

答案：【HANAstoredthemostoftenuseddatainmemory】24.單選題：Inspark,Parallelprocessingframeworkbasedonthememorycomputingmodel.ItcanbebuiltontheHadoopplatformandusetheHDFSfilesystemtostoredata,buta()architectureisbuiltontopofthefilesystemforSupportsefficientDistributedMemoryComputing.

選項(xiàng)：

A、Datachunk

B、ResilientDistributeddataset(RDD)

C、DataBlock

D、Dataset

答案：【ResilientDistributeddataset(RDD)】25.單選題：MPP(MassivelyParallelProcessing)improvesperformancethrough()parallelism.()coordinatesworkwith(),()coordinatesworkwithoneormore().()processqueriesinparallel.()havetheirownCPUdiskmemoryinsharednothingarchitecture.Highspeedinterconnectforcontinuouspipeliningofdataprocessing.

選項(xiàng)：

A、segmenthosts,Master,segmenthost,Segmenthost,segmentinstances,Segmentinstances,Segmenthosts

B、segmentinstance,Master,segmenthost,Segmenthost,segmentinstances,Segmentinstances,Segmenthosts

C、segmentinstance,Master,segmenthost,Segmenthost,segmentinstances,Segmentinstances,Segmentinstances

D、segmenthosts,Master,segmenthost,Segmenthost,segmentinstances,Segmenthosts,Segmenthosts

答案：【segmentinstance,Master,segmenthost,Segmenthost,segmentinstances,Segmentinstances,Segmenthosts】26.單選題：The()annotationtransparentlytranslatesyourPythonprogramsintoTensorFlowgraphs.

選項(xiàng)：

A、Tf.keras

B、tf.function

C、PremadeEstimators

D、tf.data

答案：【tf.function】27.單選題：（）isauser-friendlyAPIstandardformachinelearning,willbethecentralhigh-levelAPIusedtobuildandtrainmodels.

選項(xiàng)：

A、SaveModel

B、TensorFlowHub

C、PremadeEstimators

D、Tf.keras

答案：【Tf.keras】28.單選題：Inthedatastoringsystem,whichpartistheonewhichsupportCross-platformheterogeneousdatainadistributedenvironmentandprovidesthedatatoupperleveldataprocessingsystem.()

選項(xiàng)：

A、Datacollectionandmodeling

B、Distributedfilesystem

C、DistributedDatabase/datawarehouse

D、UnifiedDataAccessInterface

答案：【UnifiedDataAccessInterface】29.單選題：RelationaldatabasesandNoSQLdatabaseshavetheirownadvantagesanddisadvantagesandcannotbereplacedbyeachother,()applicationscenarios:Keybusinesssystemsintelecommunications,bankingandotherfieldsneedtoensurestrongtransactionconsistency;()applicationscenarios:Non-criticalbusiness(suchasdataanalysis)ofInternetcompanies,traditionalcompanies.

選項(xiàng)：

A、NoSQLdatabaseRelationaldatabase;

B、Relationaldatabase;NoSQLdatabase

C、NoSQLdatabase,C.NoSQLdatabase;

D、Relationaldatabase;Relationaldatabase

答案：【Relationaldatabase;NoSQLdatabase】30.單選題：In()database,therecordcan"self-describe"thetypeandcontentofthedatacontained.Thedatarecordcontainedisacollectionofaseriesofdataitems.Eachdataitemhasanameandacorrespondingvalue.Thevaluecanbeasimpledatatypesuchasastring,anumber,andadate,oritcanbeacomplextype,suchasorderedlistsandassociatedobjects.

選項(xiàng)：

A、Key-valuedatabase

B、Columnbaseddatabase

C、Document-OrientedDB

D、Graphbaseddatabase

答案：【Document-OrientedDB】31.單選題：In()database,therelationshipisdirectlydefinedandstoredinthedatabase,TraversingrelationshipisfastastheyarealreadycapturedintotheDB,andthereisnoneedtocalculatethem.

選項(xiàng)：

A、Key-valuedatabase

B、Columnbaseddatabase

C、Document-OrientedDB

D、Graphbaseddatabase

答案：【Graphbaseddatabase】32.單選題：WhichofthefollowingisNOTtheChallengestoRDBMS?()

選項(xiàng)：

A、Flexibilityandeasilyexpandedinadistributedenvironment

B、Fastquerybyindex

C、Storingunstructuredorsemi-structureddata

D、Supporttheupper-levelcalculationmodellikeDivide-and-Conquerstrategy

答案：【Fastquerybyindex】33.單選題：Whatattributessubsetselectionmethodshowedinthediagram?

選項(xiàng)：

A、ForwardStepwiseAttributessubsetselection

B、BackwardStepwiseAttributessubsetselection

C、Combineforwardselectionandbackwarddeletion

D、Decisiontree(decisiontree)induction

答案：【ForwardStepwiseAttributessubsetselection】34.單選題：Sparkadvantagesincludes()1)Fastprocessing2)Flexibility3)In-memorycomputing4)Real-timeprocessing5)Betteranalytics6)Faulttolerance7)Needextrapersistentstorage

選項(xiàng)：

A、123567

B、123456

C、134567

D、234567

答案：【123456】35.單選題：Sparkhasseveralcomponentstofacilitatedifferenttypeofcomputingtasks,likestreaming,Graphetc.thecomponentsinclude()1)SparkCoreAPI2)Resilientdistributeddataset(RDD),3）SparkSQL4）Sparktopology5）SparkStreaming6）MLlib(MachineLearningLibrary)7)GraphX8）Sklearn

選項(xiàng)：

A、12345

B、13456

C、13567

D、13578

答案：【13567】36.單選題：InStorm,()managesmanyworkernodesthroughagroupof(),Eachworkernoderunsa()daemon,monitorsthestatusofthelocalnode,andstartsandshutsdownthe()processofthenodewhennecessaryaccordingto()instructions.

選項(xiàng)：

A、Zookeeper,Nimbus,Supervisor,worker,Nimbus

B、Nimbus,Zookeeper,Supervisor,worker,Zookeeper

C、Nimbus,Zookeeper,Supervisor,worker,Nimbus

D、Supervisor,Zookeeper,Supervisor,worker,Nimbus

答案：【Nimbus,Zookeeper,Supervisor,worker,Nimbus】37.單選題：()isresponsibleforresourcemonitoringandjobscheduling,()monitorsthehealthstatusofall()andJobs,andifitfindsafailure,itwilltransferthecorrespondingtaskstoothernodes.()willtrackthetaskexecutionprogress,resourceusage,andotherinformation,andinformthe(),and()willselecttheappropriatetasktousetheseresourceswhenresourcesbecomefree.

選項(xiàng)：

A、JobTracker,JobTracker,TaskTrackers,JobTracker,TaskScheduler,TaskScheduler

B、JobTracker,TaskTrackers,JobTracker,JobTracker,TaskScheduler,TaskScheduler

C、JobTracker,JobTracker,JobTracker,TaskTrackers,TaskScheduler,TaskScheduler

D、JobTracker,JobTracker,TaskTrackers,TaskScheduler,JobTracker,TaskScheduler

答案：【JobTracker,JobTracker,TaskTrackers,JobTracker,TaskScheduler,TaskScheduler】38.單選題：WhichofthefollowingisNOTtheAdvantagesofNoSQLdatabase?

選項(xiàng)：

A、Cansupportultra-large-scaledatastorage

B、FlexibledatamodelcanwellsupportWeb2.0applications

C、Stronghorizontalexpansioncapabilities

D、Mathematicaltheoreticalfoundation

答案：【Mathematicaltheoreticalfoundation】39.單選題：MapReduceprogramwrittenbytheuserissubmittedtothe()throughthe()Userscanviewjobrunningstatusthroughsomeinterfacesprovidedby().

選項(xiàng)：

A、JobTracker,Client,Tasktracker

B、JobTracker,Client,Client

C、JobTracker,Tasktracker,Client

D、Tasktracker,Client,JobTracker

答案：【JobTracker,Client,Client】40.單選題：Classificationbelongsto()algorithmcategory.

選項(xiàng)：

A、Supervisedlearning

B、Unsupervisedlearning

C、Semi-supervisedlearning

D、Reinforcementlearning

答案：【Supervisedlearning】41.單選題：TherightorderofwritingtotheDatanodesinHDFSis()a)DistributedFileSystemmakesanRPCcalltothenamenodetocreateanewfileinthefilesystem’snamespace,withnoblocksassociatedwithit.?b)Theclientcreatesthefilebycallingcreate()methodonDistributedFileSystem.c)Thelistofdatanodesformsapipeline,anddefaultreplicationlevelisthree,sotherearethreenodesinthepipeline.TheDataStreamerstreamsthepacketstothefirstdatanodeinthepipeline,whichstoresthepacketandforwardsittotheseconddatanodeinthepipeline.d)Thenamenodeperformsvariouscheckstomakesurethefiledoesn’talreadyexistandtheclienthastherightpermissionstocreatethefile.Ifallthesecheckspass,thenamenodemakesarecordofthenewfile;otherwise,filecreationfailsandtheclientisthrownanIOException.e)TheDistributedFileSystemreturnsanFSDataOutputStreamfortheclienttostartwritingdatatodatanode.FSDataOutputStreamwrapsaDFSOutputStreamwhichhandlescommunicationwiththedatanodesandnamenode.f)Astheclientwritesdata,DFSOutputStreamsplitsitintopackets,whichitwritestoaninternalqueue,calledthedataqueue.ThedataqueueisconsumedbytheDataStreamer,whichisresponsibleforaskingthenamenodetoallocatenewblocksbypickingalistofsuitabledatanodestostorethereplicas.g)Similarly,theseconddatanodestoresthepacketandforwardsittothethird(andlast)datanodeinthepipeline.h)DFSOutputStreamalsomaintainsaninternalqueueofpacketsthatarewaitingtobeacknowledgedbydatanodes,calledtheackqueue.Apacketisremovedfromtheackqueueonlywhenithasbeenacknowledgedbyallthedatanodesinthepipeline.i)Whentheclienthasfinishedwritingdata,itcallsclose()onthestream.Itflushesalltheremainingpacketstothedatanodepipelineandwaitsforacknowledgmentsbeforecontactingthenamenodetosignalthatthefileiscomplete.Thenamenodealreadyknowswhichblocksthefileismadeupof,soitonlyhastowaitforblockstobeminimallyreplicatedbeforereturningsuccessfully.

選項(xiàng)：

A、badefcghij

B、abdefcghij

C、badefchgij

D、badefcgihj

答案：【badefcghij】42.單選題：AboutDataModelingdesignlevelsdescriptions:Whichoneiscorrectmatching?()1)Basedontheuser'sdatafunctionrequirements.functionsandassociationrelationshipsareobtained,EntityClasscorrespondingtothebusinesselementsandfunctions.2)Moredetailsofdataentities,includingprimarykeys,foreignkeys,attributes,indexes,relationships,constraints,andevenviews,withdatatables,datacolumns,valueranges,object-orientedclasses,XMLtagsandotherformstodescribe.3)Thestorageimplementationofdata,includingdatapartition,datatablespace,anddataintegration.

選項(xiàng)：

A、1-Conceptualmodeldesign2-physicalmodeldesign3-logicalmodeldesign

B、1-Logicalmodeldesign2-Physicalmodeldesign3-Conceptualmodeldesign

C、1-Conceptualmodeldesign2-logicalmodeldesign3-Physicalmodeldesign

D、1-Physicalmodeldesign2-Conceptualmodeldesign3-logicalmodeldesign

答案：【1-Conceptualmodeldesign2-logicalmodeldesign3-Physicalmodeldesign】43.單選題：Basedontherequirementwecanbuildthebusinessmodel,itincludes()and().

選項(xiàng)：

A、ConceptualModel,LogicModel

B、LogicModel,PhysicalModel

C、Processmodel,Datamodel

D、Processmodel,logicalmodel

答案：【Processmodel,Datamodel】44.單選題：WhichofthefollowingisnotcorrectaboutNoSQLDatabasestructurecharacters?

選項(xiàng)：

A、Nopre-defineddataformatisrequired;

B、Flexibleandscalable;

C、Applysharedstoragearchitecture;

D、Asynchronousreplication;

答案：【Applysharedstoragearchitecture;】45.單選題：TensorFlowallowsdeveloperstocreate()—structuresthatdescribehowdatamovesthrougha(),oraseriesofprocessingnodes.Eachnodeinthegraphrepresentsa(),EachconnectionorEdgebetweennodesisa().

選項(xiàng)：

A、DataflowGraphs,Graph(DAG),multidimensionaldataarrayortensor,mathematicaloperation

B、Graph(DAG),DataflowGraphs,mathematicaloperation,multidimensionaldataarrayortensor

C、DataflowGraphs,Graph(DAG),mathematicaloperation,multidimensionaldataarrayortensor

D、Graph(DAG),DataflowGraphs,mathematicaloperation,multidimensionaldataarrayortensor

答案：【DataflowGraphs,Graph(DAG),mathematicaloperation,multidimensionaldataarrayortensor】46.單選題：RepresentativecomputingplatformsofBatchProcessingModel;StreamComputingmodel;MPP,massiveparallelprocessing;In-memorycomputing;GraphParallelModelsare()

選項(xiàng)：

A、Hadoop,Greenplum,Storm,Spark,Pregel

B、Hadoop,Storm,Greenplum,Spark,Pregel

C、Hadoop,Storm,Pregel,Spark,Greenplum

D、Hadoop,Spark,Greenplum,Storm,Pregel

答案：【Hadoop,Storm,Greenplum,Spark,Pregel】47.單選題：Self-trainingbelongsto()algorithmcategory.

選項(xiàng)：

A、Supervisedlearning

B、Unsupervisedlearning

C、Semi-supervisedlearning

D、Reinforcementlearning

答案：【Semi-supervisedlearning】48.單選題：Clusteringbelongsto()algorithmcategory.

選項(xiàng)：

A、Supervisedlearning

B、Unsupervisedlearning

C、Semi-supervisedlearning

D、Reinforcementlearning

答案：【Unsupervisedlearning】49.單選題：Inthefollowing,whichonecanachievelinearperformancescalingwhenCPUincreases?

選項(xiàng)：

A、SMP

B、NUMA

C、MPP

D、Noneofthem

答案：【MPP】50.單選題：TheFeaturesofSparkinclude()1)In-memoryComputation2)On-diskComputation3)LazyEvaluation4)ImmediatelyEvaluation5)FaultTolerant6)Immutability7)Partitioning8)Persistence9)Coarse-GrainedOperation10)Fine-GrainedOperation

選項(xiàng)：

A、2346789

B、1356789

C、135678,10

D、235678,10

答案：【1356789】51.單選題：InSpark，()isresponsibleforconvertingthecomputingtasksoftheapplicationintoa().()isresponsibleforcompletingthecalculationanddatastorageontheworkernode.Oneachworker,the()generatestaskthreadsforeachdatapartitiondistributedtoittocompleteparallelcalculations.

選項(xiàng)：

A、Executor,topology，Executor,Driver

B、Driver,Directedacyclicgraph(DAG)，Executor,Executor

C、Executor,Directedacyclicgraph(DAG)，Executor,Driver

D、Driver,topology，Executor,Executor

答案：【Driver,Directedacyclicgraph(DAG)，Executor,Executor】52.單選題：Theprocessbeginsbythe()issuingaquerythatisthenpassedtothe().The()containsinformation,suchasthedatadictionaryandsessioninformation,whichitusestogeneratean()designedtoretrievetheneededinformationfromeachunderlyingNode.ParallelExecutionrepresentstheimplementationofthe()throughtheparallelcomputingofNode1toNoden.Andthequeryresultsreturntomasternode.

選項(xiàng)：

A、Client,MasterNode,MasterNode,executionplan,executionplan

B、MasterNode,Client,MasterNode,executionplan,storingplan

C、Client,MasterNode,Client,executionplan,executionplan

D、MasterNode,Client,MasterNode,executionplan,storingplan

答案：【Client,MasterNode,MasterNode,executionplan,executionplan】53.單選題：ThecorrectdescriptionsofcharacteristicsandrequirementoftheStreamComputinginclude()1)Dataisnolongerarrivinginbatchesbutcontinuouslyarrivingdynamically2)Computationalanalysisrequiresreal-time,fastresponse,andlowlatency3)Theamountofdataislarge,thestorageofthedataisnotvalued,buttheimmediateprocessingandanalysisofthedataareemphasized4)Payattentiontothecalculationandanalysisresultsofthedataasawhole,butnottopayattentiontotheindividualdata5)Theorderandtimingofthearrivalofdataelementscannotbepredictedorcontrolled,andthecalculationprogrammustbeabletorespond6)real-timeanalysisandcalculationfordynamiccontinuousdatastreams7)Afterthecalculationresultsareobtained,thedataiseitherimportedintothestaticdatabase,eitherdiscard,thatis,one-timeuse.

選項(xiàng)：

A、12347

B、1234567

C、124567

D、123567

答案：【1234567】54.單選題：AboutthePrincipalcomponentanalysis(PCA),whichstatementiswrong?

選項(xiàng)：

A、Principalcomponentanalysissearchstoobtainc-dimensionalorthogonalvectorsthatbestrepresentthedata.

B、PrincipalcomponentanalysisistheNumerosityreductionmethods.

C、Theoriginaldatacanbeprojectedintoasmallerspacetoachievedatacompression.

D、Principalcomponentanalysisislossycompression.

答案：【PrincipalcomponentanalysisistheNumerosityreductionmethods.】55.單選題：InHDFS,thenamenodeandthedatanodehavetheirownresponsibilities,selecttheresponsibilitiesofnamenodeanddatanoderespectively.Namenodes(),Datanodes()1)Realizethemappingofdatablockstothelocalfilesystemofthedatanode2)Managefilesystemnamespace3)Storefiledatablock4)Save“filetodatablocktodatanode”mappingrelationship5)Schedulingclientaccesstofiles6)storetheDatablocksonthelocaldisk7)StoretheMetadatainmemoryforquickaccess

選項(xiàng)：

A、1237,456

B、2457,136

C、1245,367

D、2456,137

答案：【2457,136】56.單選題：WhichoneofthefollowingisnotthecharactersofNoSQLDatabasetransactionmodel–BASE?

選項(xiàng)：

A、BasicallyAvailable

B、Eventualconsistency

C、Softstate

D、Atomicity

答案：【Atomicity】57.單選題：AccordingtotheCAPtheorem,adistributedsystemcanonlymeettwooftheCAPsinthemid-rundatareadandwriteoperations,whichpropertiescanRDBMSsatisfy?

選項(xiàng)：

A、consistency(C)andavailability(A)

B、consistency(C)andpartitiontolerance(P)

C、availability(A)andpartitiontolerance(P)

D、Noneofthem

答案：【consistency(C)andavailability(A)】58.單選題：WhichofthefollowingisNOTcorrectdescriptionofNoSQLdatabase?()

選項(xiàng)：

A、NoSQLisSchemaless

B、Partitionusinghashandkeyspace

C、NoSQLdealawaywithcostlyrelationship,standonitsown.socanscalewell

D、NoSQLcanhavethehardconsistency

答案：【NoSQLcanhavethehardconsisten