简介：Thispaperintegratesgeneticalgorithmandneuralnetworktechniquestobuildnewtemporalpredictinganalysistoolsforgeographicinformationsystem(GIS).ThesenewGIStoolscanbereadilyappliedinapracticalandappropriatemannerinspatialandtemporalresearchtopatchthegapsinGISdataminingandknowledgediscoveryfunctions.ThespecificachievementhereistheintegrationofrelatedartificialintelligenttechnologiesintoGISsoftwaretoestablishaconceptualspatialandtemporalanalysisframework.And,byusingthisframeworktodevelopanartificialintelligentspatialandtemporalinformationanalyst(ASIA)systemwhichthenisfullyutilizedintheexistingGISpackage.Thisstudyofairpollutantsforecastingprovidesageographicalpracticalcasetoprovetherationalizationandjustnessoftheconceptualtemporalanalysisframework.

简介：Inthispaper,weinvestigatecoherenceresonance(CR)andnoise-inducedsynchronizationinHindmarsh-Rose(HR)neuralnetworkwiththreedifferenttypesoftopologies:regular,random,andsmall-world.ItisfoundthattheadditivenoisecaninduceCRinHRneuralnetworkwithdifferenttopologiesanditscoherenceisoptimizedbyapropernoiselevel.Itisalsofoundthatascouplingstrengthincreasestheplateauinthemeasureofcoherencecurvebecomesbroadenedandtheeffectsofnetworktopologyismorepronouncedsimultaneously.Moreover,wefindthatincreasingtheprobabilitypofthenetworktopologyleadstoanenhancementofnoise-inducedsynchronizationinHRneuronsnetwork.

简介：BACKGROUND:Exposuretolow-levelleadhasatoxiceffectonthedevelopmentofneonates,whichhasattractedwideattention.Colostrumleadlevelcanbeusedastheindicationofleadexposure.OBJECTIVE:Toobservetherelationshipofcolostrumleadlevelandtheneurobehavioraldevelopmentofinfants.DESIGN:Aprospectivecontrolobservation.SETTING:CenterforMaternalandChildHealth,ShanxiProvincialChildren'sHospital.PARTICIPANTS:Totally128neonatesoffull-termnormaldelivery,76maleand52female,fromShanxiProvincialMaternalandChildHealthCenterandJiexiuMaternalandChildHealthCenterwereinvolvedinthisstudy.Alltheinvolvedneonateshadnoperipartalischemic/hypoxichistoryorfetusintrauterinedevelopmentallag.Pregnantwomenhadnovariousacuteandchronicdiseasesinpregnancy,familyhistoryofneurologicaldiseaseoroccupationalleadexposure.128portionsofcolostrumsampleoffull-termnormaldeliverywerecollected.Informedconsentsofdetecteditemswereobtainedfromthepuerperantsandtheirrelatives.METHODS:①Experimentalgrouping:Leadlevelinthecolostrumwasdeterminedbyatomicabsorptionspectrometry.Accordingtoleadlevelinthecolostrum,theneonateswereclassifiedintotwoexposuregroupsofgreaterthanorequalto0.24μmol/Linahigh-levelleadgroupandlessthan0.24μmol/Linalow-levelleadgroup.②Experimentalevaluation:Mentaldevelopmentalindex(MDI)andpsychologicaldevelopmentalindex(PDI)of3-month-oldinfantswereevaluatedwithBayleyScalesofInfantDevelopment(BSID).TherelationshipsofMDI,PDIandcolostrumleadlevelwereperformedcorrelationregressionanalysis;Therelationshipofcolostrumleadlevelanddevelopmentwasperformedmulti-factoranalysiswithfamilyenvironmentandhealthquestionnaires.MAINOUTCOMEMEASURES:①EvaluationresultsofMDIandPDI.②Multi-factoranalysisresults.RESULTS:Totally128neonateswereinvolvedinthestudy.Tenandelevenneonateswerelostduetoemigra

简介：Laserblankweldingisbecomingmoreandmoreimportantintheautomotiveindustryandthequalityoftheweldiscriticalforasuccessfulapplication.Afullyautomatedsolutionisrequiredtoinspectthequalityoftheblanks.ThispaperpresentsavisioninspectionsystemwithaCMOScamerawhichusesART2networktoinspectthedefectson-linetoobtainthegeometryandthequalityoftheweldseam.TheneuralnetworkART2hasthecapabilityofself-learningfromtheenvironment.Itcandistinguishthedefectsthathavebeenlearnedbeforeandgivenewoutputsfornewdefects.SoART2networkissuitableforweldqualityinspectioninlaserblankwelding.Additionally,aCO_2laserisusedfortheblankbutt-welding.

简介：Therelationbetweenthewaterdischarge(Q)andsuspendedsedimentconcentration(SSC)oftheRiverRamgangaatBareilly,UttarPradesh,intheHimalayas,hasbeenmodeledusingArtificialNeuralNetworks(ANNs).Thecurrentstudyvalidatesthepracticalcapabilityandusefulnessofthistoolforsimulatingcomplexnonlinear,realworld,riversystemprocessesintheHimalayanscenario.ThemodelingapproachisbasedonthetimeseriesdatacollectedfromJanuarytoDecember(2008-2010)forQandSSC.ThreeANNs(T1-T3)withdifferentnetworkconfigurationshavebeendevelopedandtrainedusingtheLevenbergMarquardtBackPropagationAlgorithmintheMatlabroutines.Networkswereoptimizedusingtheenumerationtechnique,and,finally,thebestnetworkisusedtopredicttheSSCvaluesfortheyear2011.ThevaluesthusobtainedthroughtheANNmodelarecomparedwiththeobservedvaluesofSSC.Thecoefficientofdetermination(R2),fortheoptimalnetworkwasfoundtobe0.99.ThestudynotonlyprovidesinsightintoANNmodelingintheHimalayanriverscenario,butitalsofocusesontheimportanceofunderstandingariverbasinandthefactorsthataffecttheSSC,beforeattemptingtomodelit.Despitethetemporalvariationsinthestudyarea,itispossibletomodelandsuccessfullypredicttheSSCvalueswithverysimplisticANNmodels.

简介：Thejoiningofa6-mmthicknessAl6061toStainlesssteel304hasbeenperformedbysolidstatewelding.Aselectionmethodofoptimumfrictionweldingconditionusingneuralnetworksisproposed.Thedatausedforanalysesarethefrictionstirweldingcondition,theinputparametersofthemodelconsistofweldingspeedandtoolrotationspeed.TheoutputsoftheANN(ArtificialNeuralNetwork)modelincludesresultingparameters,namely,maximumreachedtemperature,andheatingrateforbothaluminumalloy6061andstainlesssteel304duringfrictionstirweldingprocess.Theresultsofanalysissuggestthattheproposedmethodisaneffectiveonetoselectanoptimumweldingcondition.GoodperformanceoftheANNmodelwasachieved.Thecombinedinfluenceofweldingspeedandtoolrotationspeedonthemaximumreachedtemperatureandheatingrateforbothaluminumalloy6061andstainlesssteel304frictionstirweldingwassimulated.AcomparisonwasmadebetweentheoutputoftheANNprogramandfiniteelementmodel.Thecalculatedresultswereingoodagreementwiththatoffiniteelementmodel.

简介：Automobilecompaniesthatspendbillionsofdollarsannuallytowardswarrantycost,givehighprioritytowarrantyreductionprograms.Forecastingofautomobilewarrantyperformanceplaysanimportantroletowardstheseefforts.Theforecastingprocessinvolvespredictionofnotonlythespecificmonths-in-service(MIS)warrantyperformanceatcertainfuturetime,butalsoatfutureMISvalues.However,'maturingdata'(alsocalledwarrantygrowth)phenomenathatcauseswarrantyperformanceatspecificMISvaluestochangewithtime,makessuchaforecastingtaskchallenging.Althoughwarrantyforecastingmethodssuchaslog-logplotsanddynamiclinearmodelsappearinthispaperweuseanartificialneuralnetworkfortheforecastingofwarrantyperformanceinpresencetestingerrorsusingresponsesurfacemethodology.Thisapplicationshowstheeffectivenessofneuralphenomena.

简介：传统地，在空中的时间域电磁(ATEM)数据被转换由重复导出地球模型。然而，数据高度经常在隧道之中被相关并且因而在倒置引起提出病、在坚定上的问题。关联复杂化在ATEM数据和地球参数之间的印射的关系并且因此增加倒置复杂性。排除这，我们采用主要部件分析把ATEM数据转变成直角的主要部件(PC)减少关联和数据维数并且同时压制无关的噪音。在这份报纸，我们使用一个人工的神经网络(ANN)接近与地球模型参数印射关系的PC，避免Jacobian衍生物的计算。基于PC的ANN算法为在空中的时间域与data-basedANN相比为分层的模型被用于合成数据电磁的倒置。结果在data-basedANN上表明更简单的网络结构，更少的训练步骤，和更好的倒置结果的基于PC的ANN优点，特别为污染数据。而且，基于PC的ANN算法有效性被假2D模型和比较的倒置与data-basedANN和Zhodys方法检验。结果显示基于PC的ANN倒置能与真正的模型一起完成一个更好的协议并且也证明基于PC的ANN是可行的转换大ATEM数据集。

简介：TheauthorsdiscussproblemsofapproximationtofunctionsinL^2(R^n)andoperatorsfromL^2(R^n1)toL^2(R^n2)byRadial-BasisFunctions.TheresultsobtainedsolvetheparblemofcapabilityofRBFneuralnetworks,abasicprobleminneuralnetworks.

简介：当前的纸在卡梅伦区域附近论述山崩危险分析，马来西亚，用在地理信息系统(GIS)和遥感的帮助下的先进人工的神经网络技术。山崩地点被天线相片并且从领域调查的解释在学习区域决定。地形学、地质的数据以及卫星图象被收集，处理，并且构造进用GIS并且图象处理的一个空间数据库。十个因素为山崩危险包括被选择：1)与是的地形学有关的因素斜坡，方面，和弯曲；2)与地质学有关的因素作为从貌的岩性学和距离；3)与排水有关的因素作为从排水的距离；并且4)作为陆地盖子和植被索引从TM卫星图象提取的因素珍视。一个先进人工的神经网络模型被用来分析这些因素以便建立山崩危险地图。训练方法的背繁殖被用于五个不同随机的训练地点的选择以便计算因素重量然后山崩危险索引为每五张危险地图被计算。最后，山崩危险地图(五个盒子)用GIS工具被准备。山崩危险地图的结果用山崩测试地点被验证了没在神经网络的训练阶段期间被使用。我们确认的调查结果结果为训练地点显示出69%，75%，70%，83%和86%的精确性1，2，3，4和5分别地。GIS数据被用来高效地分析大量数据，并且人工的神经网络证明了是为山崩危险分析的一个有效工具。确认结果在山崩区域上显示出在假定危险地图和存在数据之间的足够的同意。

简介：

简介：Thealternatecombinationalapproachofgeneticalgorithmandneuralnetwork(AGANN)hasbeenpresentedtocorrectthesystematicerrorofthedensityfunctionaltheory(DFT)calculation.IttreatstheDFTasablackboxandmodelstheerrorthroughexternalstatisticalinformation.Asademonstration,theAGANNmethodhasbeenappliedinthecorrectionofthelatticeenergiesfromtheDFTcalculationfor72metalhalidesandhydrides.ThroughtheAGANNcorrection,themeanabsolutevalueoftherelativeerrorsofthecalculatedlatticeenergiestotheexperimentalvaluesdecreasesfrom4.93%to1.20%inthetestingset.Forcomparison,theneuralnetworkapproachreducesthemeanvalueto2.56%.Andforthecommoncombinationalapproachofgeneticalgorithmandneuralnetwork,thevaluedropsto2.15%.Themultiplelinearregressionmethodalmosthasnocorrectioneffecthere.

简介：Duetothedemandofdataprocessingforpolariceradarinourlaboratory,aCurveletThresholdingNeuralNetwork(TNN)noisereductionmethodisproposed,andanewthresholdfunctionwithinfinite-ordercontinuousderivativeisconstructed.ThemethodisbasedonTNNmodel.InthelearningprocessofTNN,thegradientdescentmethodisadoptedtosolvetheadaptiveoptimalthresholdsofdifferentscalesanddirectionsinCurveletdomain,andtoachieveanoptimalmeansquareerrorperformance.Inthispaper,thespecificimplementationstepsarepresented,andthesuperiorityofthismethodisverifiedbysimulation.Finally,theproposedmethodisusedtoprocesstheiceradardataobtainedduringthe28thChineseNationalAntarcticResearchExpeditionintheregionofZhongshanStation,Antarctica.Experimentalresultsshowthattheproposedmethodcanreducethenoiseeffectively,whilepreservingtheedgeoftheicelayers.

简介：Objective:Toevaluatetheeffectsofhyperbaricoxygen(HBO)therapyonpatientswithpostbraininjuryneuralstatus.Methods:Twoto4coursesofHBOtherapyand/ormedicationswereusedtotreat320patientswhowererandomlydividedintotwogroups.Assessmentwasmadewith99mTc-ethylcysteinatedimer(99mTc-ECD)singlephotonemissioncomputedtomography(SPECT)beforeandaftertreatment.Results:TherewasasignificantdifferencebetweentheHBOtherapygroupandthenon-HBOtherapygroup.HBOtherapywassuperiortomedicationtreatmentaloneintherecoveryofclinicalsymptoms,controlofepilepsy,andresolutionofhydrocephalus(P<0.01).Conclusions:HBOtherapyhasspecificcurativeeffectsonpatientswithpostbraininjuryneuralstatus,and99mTc-ECDSPECTcouldplayanimportantroleindiagnosingpostbraininjuryneuralstatusandmonitoringthetherapeuticeffectsofHBO.

简介：BasedonthefuzzycharacteristicofthepulsestateandsyndromesdifferentiationthinkingmodeofTCM,aninformationfusingrecognitionmethodofpulsestatesbasedonSFNN(StochasticFuzzyNeuralNetwork)ispresentedinthispaper.Withthelearningabilityinparametersandstructure,SFNNfusesthemeasurementinformationofthreepulse-statesensorsdistributedinCun,Guan,andChilocationofbodyforthepulsestaterecognition.Theexperimentalresultsshowthatthepercentageofcorrectrecognitionwithnewmethodishigherthanthatbysingle-datarecognitionone,withfeweroff-linetrainnumbers.

简介：ＭｏｎｔｈｌｙＭｅａｎＴｅｍｐｅｒａｔｕｒｅＰｒｅｄｉｃｔｉｏｎＢａｓｅｄｏｎａ　Ｍｕｌｔｉ－ｌｅｖｅｌＭａｐｐｉｎｇＭｏｄｅｌｏｆＮｅｕｒａｌＮｅｔｗｏｒｋＢＰＴｙｐｅＹａｎＳｈａｏｊｉｎ；（严绍瑾），ＰｅｎｇＹｏｎｇｑｉｎｇ；（彭永清）ａｎｄＧｕ...

简介：Anew3Dsurfacecontouringandrangingsystembasedondigitalfringeprojectionandphaseshiftingtechniqueispresented.Usingthephase-shifttechnique,pointscloudwithhighspatialresolutionandlimitedaccuracycanbegenerated.Stereo-pairimagesobtainedfromtwocamerascanbeusedtocompute3Dworldcoordinatesofapointusingtraditionalactivetriangulationapproach,yetthecameracalibrationiscrucial.Neuralnetworkisawell-knownapproachtoapproximateanonlinearsystemwithoutanexplicitphysicalmodel,inthisworkitisusedtotrainthestereovisionapplicationsystemtocalculating3Dworldcoordinatessuchthatthecameracalibrationcanbebypassed.Thetrainingsetforneuralnetworkconsistsofavarietyofstereo-pairimagesandthecorresponding3Dworldcoordinates.Thepictureelementscorrespondenceproblemissolvedbyusingprojectedcolor-codedfringeswithdifferentorientations.Colorimbalanceiscompletelyeliminatedbythenewcolor-codedmethod.Oncethehighaccuracycorrespondenceof2Dimageswith3Dpointsisacquired,highprecision3Dpointscloudcanberecognizedbythewelltrainednet.Theobviousadvantageofthisapproachisthathighspatialresolutioncanbeobtainedbythephase-shiftingtechniqueandhighaccuracy3Dobjectpointcoordinatesareachievedbythewelltrainednetwhichisindependentofthecameramodelworksforanytypeofcamera.Someexperimentsverifiedtheperformanceofthemethod.

简介：Thegeophysicalmodelfunction(GMF)describestherelationshipbetweenbackscatteringandseasurfacewind,sothatwindvectorscanberetrievedfrombackscatteringmeasurement.TheGMFplaysanimportantroleinoceanwindvectorretrievals,itsperformancewilldirectlyinfluencetheaccuracyoftheretrievedwindvector.Neuralnetwork(NN)approachisusedtodevelopaunifiedGMFforC-bandandKu-band(NN-GMF).EmpiricalGMFCMOD4andQSCAT-1areusedtogeneratethesimulatedtrainingdata-set,andGaussiannoiseatasignalnoiseratioof30dBisaddedtothedata-settosimulatethenoiseinthebackscatteringmeasurement.TheNN-GMFemploysradiofrequencyasanadditionalparameter,soitcanbeappliedforbothC-bandandKu-band.Analysesshowthattheσ0predictedbytheNN-GMFiscomparablewiththeσ0predictedbyCMOD4andQSCAT-1.AlsothewindvectorsretrievedfromtheNN-GMFandempiricalGMFCMOD4andQSCAT-1arecomparable,indicatingthattheNN-GMFisaseffectiveastheempiricalGMF,andhastheadvantagesoftheuniversalform.

简介：由使用Mawhins巧合度理论和Liapunov功能的方法的继续定理，一些足够的条件被获得保证存在，唯一和包含timevarying延期的欺骗类型Cohen-Grossberg神经网络的周期的答案的全球指数的稳定性。

简介：Objective:Toidentifyandseparatetheventralrootfromdorsalroot,whichisthekeyforsuccessoftheartificialsomatic-autonomicreflexpathwayprocedureforneurogenicbladderafterspinalcordinjury(SCI).Herewereporttheresultsofintra-operatingroommonitoringwith10paralyzedpatients.Methods:TenmalevolunteerswithcompletesuprasacralSCIunderwenttheartificialsomatic-autonomicprocedureundergeneralanesthesia.Vastusmedialis,tibialisanticusandgastrocnemiusmedialisoftheleftlowerlimbweremonitoredforelectromyogram(EMG)activitiesresultedfromL4,L5,andS1stimulationrespectivelytodifferentiatetheventralrootfromdorsalroot.ALaborieUrodynamicssystemwasconnectedwithathreechannelurodynamiccatheterinsertedintothebladder.TheL2andL3rootswerestimulatedseparatelywhiletheintravesicalpressurewasmonitoredtoevaluatethefunctionofeachroot.Results:Thethresholdsofstimulationonventralrootwere0.02msduration,0.2-0.4mA,(mean0.3mA±0.07mA),comparedwith0.2-0.4msduration,1.5-3mA(mean2.3mA±0.5mA)fordorsalroot(P<0.01)tocauserevokedpotentialsandEMG.ElectricalstimulationonL4rootsresultedintheEMGbeingrecordedmainlyonvastusmedialis,whilestimulationonL5orS1rootscausedelectricalactivitiesoftibialisanticusorgastrocnemiusmedialisrespectively.Thecontinuousstimulationforabout3-5secondsonS2orS3ventralroot(0.02ms,20Hz,and0.4mA)couldresultedinbladderdetrusorcontraction,butthestrongestbladdercontractionover50cmH2OwasusuallycausedbystimulationonS3ventralrootin7ofthe10patients.Conclusions:Intra-operatingroomelectrophysiologicalmonitoringisofgreathelptoidentifyandseparateventralrootfromdorsalroot,andtoselecttheappropriatesacralventralrootforbestbladderreinnervation.Differentparametersandthresholdsondifferentrootsarethemostimportantfactorstokeepinmindtoavoiddamagingtherootsandtoassurethebestresu