简介：Intrusiondetectioncanbeessentiallyregardedasaclassificationproblem,namely,dis-tinguishingnormalprofilesfromintrusivebehaviors.Thispaperintroducesboostingclassificationalgorithmintotheareaofintrusiondetectiontolearnattacksignatures.Decisiontreealgorithmisusedassimplebaselearnerofboostingalgorithm.Furthermore,thispaperemploysthePrincipleCom-ponentAnalysis(PCA)approach,aneffectivedatareductionapproach,toextractthekeyattributesetfromtheoriginalhigh-dimensionalnetworktrafficdata.KDDCUP99datasetisusedintheseex-perimentstodemonstratethatboostingalgorithmcangreatlyimprovetheclassificationaccuracyofweaklearnersbycombininganumberofsimple“weaklearners”.Inourexperiments,theerrorrateoftrainingphaseofboostingalgorithmisreducedfrom30.2%to8%after10iterations.Besides,thispaperalsocomparesboostingalgorithmwithSupportVectorMachine(SVM)algorithmandshowsthattheclassificationaccuracyofboostingalgorithmislittlebetterthanSVMalgorithm’s.However,thegeneralizationabilityofSVMalgorithmisbetterthanboostingalgorithm.

简介：Ahybridapproachforfuzzysystemdesignbasedonclusteringandakindofneurofuzzynetworksisproposed.Anunsupervisedclusteringtechniqueisfirstlyusedtodeterminethenumberofif-thenfuzzyrulesandgenerateaninitialfuzzyrulebasefromthegiveninput-outputdata.Then,aclassofneurofuzzynetworksisconstructedanditsweightsaretunedsothattheobtainedfuzzyrulebasehasahighaccuracy.Finally,twoexamplesoffunctionapproximationproblemsaregiventoillustratetheeffectivenessoftheproposedapproach.

简介：AnapproachofdefendingagainstDistributedDenialofService(DDoS)attackbasedonflowmodelandflowdetectionispresented.TheproposedapproachcanprotecttargetsfromDDoSattacking,andallowtargetstoprovidegoodservicetolegitimatetrafficunderDDoSattacking,withfastreaction.Thisapproachadoptsthetechniqueofdynamiccombfilter,yieldsalowleveloffalsepositivesoflessthan1.5%,dropssimilarpercentageofgoodtraffic,about1%,andpassesneglectablepercentageofattackbandwidthtothevictim,lessthan1.5%.Theprototypeofcommercialproduct,D-fighter,isdevelopedbyimplementingthisproposedapproachonIntelnetworkprocessorplatformIXP1200.

简介：Specklefilteringofsyntheticapertureradar(SAR)imageswhilepreservingthespatialsignalvariability(textureandfinestructures)stillremainsachallenge.ManyalgorithmshavebeenproposedfortheSARimagerydespeckling.However,simulatedannealing(SA)methodisoneofexcellentchoicescurrently.AcriticalprobleminthestudyonSAistoprovideappropriatecoolingschedulesthatensurefastconvergencetonear-optimalsolutions.Thispapergivesanewnecessaryandsufficientconditionforthecoolingschedulesothatthealgorithmstateconvergesinallprobabilitytothesetofgloballyminimumcoststates.Moreover,itconstructsanappropriateobjectivefunctionforSARimagedespeckling.AnexperimentalresultoftheactualSARimageprocessingisobtained.

简介：Thispaperdescribes,bymeansofaVoronoihypersphere,thenearestneighborrelationsofallthefeaturesubmatricesinthefaultclassificationspaceandanalysesthedeviationdisturbanceanglesbetweenfaultfeaturesubmatricesandak-dimensionunitarymatrixofthemeasuredvoltagechangematrix.Withtheabovetwoconcepts,thispaperdiscussesthediagnos-abilityinthefaultclassificationapproach.Thepaperalsoclassifiesanddefinesthefaultdiagnosisproblems.Finally,thepaperderivesthecorrespondingnecessaryandsufficientconditionsforcorrectlocationoffaults.

简介：ＡｎＡｐｐｒｏａｃｈｔｏＲＣＳＣｏｍｐｕｔａｔｉｏｎｆｏｒＣｏｍｐｌｅｘＴａｒｇｅｔｓＺｈａｏＷｅｉｊｉａｎｇ＆ＷａｎｇＭａｏｇｕａｎｇＸｉｄｉａｎＵｎｉｖｅｒｓｉｔｙ，Ｘｉ’ａｎ７１００７１，Ｐ．Ｒ．Ｃｈｉｎａ（ＲｅｃｅｉｖｅｄＡｕｇｕｓｔ９，１９...

简介：Inaerialrobots’visualnavigation,itisessentialyetverydiffculttodetecttheattitudeandpositionoftherobotsoperatedinrealtime.Byintroducinganewparametricmodel,theproblemcanbereducedfromalmostunmanageabletobepartlysolved,thoughnotfully,aspertherequirement.Inthisparametricapproach,amulti-scaleleastsquaremethodisformulatedfirst.Bypropagatingaswellasimprovingtheparametersdownfromlayertolayeroftheimagepyramid,anewglobalfeaturelinecanthenbedetectedtoparameterizetheattitudeoftherobots.Furthermore,thisapproachpavesthewayforsegmentingtheimageintodistinctparts,whichcanberealizedbydeployingaBayesianclassifieronthepicturecelllevel.ComparisonwiththeHoughtransformbasedmethodintermsofrobustnessandprecisionshowsthatthismulti-scaleleastsquarealgorithmisconsiderablymorerobusttonoises.Somediscussionsarealsogiven.

简介：AdmissioncontrolplaysanimportantroleinprovidingQoStonetworkusers.Motivatedbythemeasurement-basedadmissioncontrolalgorithm,thisletterproposedanewadmissioncontrolapproachforintegratedservicepacketnetworkbasedontrafficprediction.Intheletter,FARIMA(p,d,q)modelsintheadmissioncontrolalgorithmisdeployed.AmethodtosimplifytheFARIMAmodelfittingprocedureandhencetoreducethetimeoftrafficmodelingandpredictionissuggested.Thefeasibility-studyexperimentsshowthatFARIMAmodelswhichhavelessnumberofparameterscanbeusedtomodelandpredictactualtrafficonquitealargetimescale.Simulationresultsvalidatethepromisingapproach.

简介：ThispaperpresentsapragmaticadaptiveschemeforTuCMoverslowlyfadingchannels.Theadaptiveschemeemploysasingleturbocodedmodulatorcomposedofavariable-rateturboencoderandavariable-ratevariable-powerMQAMforallfadingregions,soithasanacceptablecomplexitytoimplement.TheoptimaladaptiveTuCMschemeisdeterminedsubjecttovarioussystemconstraints.Simulationshavebeenperformedtomeasuretheperformanceoftheschemefordifferentparameters.Itisshownthatadoptingboththeturbocodedmodulatorandthetransmitpowerachievesaperformancewithin2.5dBofthefadingchannelcapacity.

简介：Gatematrixlayoutproblemplaysanimportantroleinintegratedcircuitdesign,butitsoptimizationisNP-hard.Inthispaper,typicalgatelayoutproblemisanalysedandadaptedtoneuralnetworkrepresentation,furthermorethesimulatedresultsaregiven.

简介：ＡｎＡｐｐｒｏａｃｈｆｏｒＩｎｔｅｇｒａｔｉｎｇＱｕａｎｔｉｔａｔｉｖｅＤｅｉｓｉｏｎＭｏｄｅｌｗｉｔｈＱｕａｌｉｔａｔｉｖｅＪｕｄｇｍｅｎｔTXＡｎＡｐｐｒｏａｃｈｆｏｒＩｎｔｅｇｒａｔｉｎｇＱｕａｎｔｉｔａｔｉｖｅＤｅｃｉｓｉｏｎＭｏｄｅｌｗｉｔｈ...

简介：侦察混血儿雷达信号的一条参数评价途径联合了调音的频率移动(频移键控)，调音的阶段移动(PSK)被介绍。第一，多相的差别被采用计算即时频率(如果)FSK/PSK，那么，频移键控的频率点从直方图被估计如果。PSK的代码率从阶段断绝的地点被提取。最后，收到的信号的平方的多相的差别被计算估计频移键控的代码率。当signal-to-noise比率(SNR)在11dB上面时，介绍算法有参数评价的更高的精确性。

简介：Namedentityrecognitionisafundamentaltaskinbiomedicaldatamining.Inthisletter,anamedentityrecognitionsystembasedonCRFs(ConditionalRandomFields)forbiomedicaltextsispresented.Thesystemmakesextensiveuseofadiversesetoffeatures,includinglocalfeatures,fulltextfeaturesandexternalresourcefeatures.Allfeaturesincorporatedinthissystemaredescribedindetail,andtheimpactsofdifferentfeaturesetsontheperformanceofthesystemareevaluated.Inordertoimprovetheperformanceofsystem,post-processingmodulesareexploitedtodealwiththeabbrevia-tionphenomena,cascadednamedentityandboundaryerrorsidentification.Evaluationonthissystemprovedthatthefeatureselectionhasimportantimpactonthesystemperformance,andthepost-processingexploredhasanimportantcontributiononsystemperformancetoachievebetterre-sults.

简介：信息熔化是在多模式的生物测定的系统的关键步。特征级的熔化是比由于原来的特征集合关于生物测定的数据包含更富有的信息的事实的分数级、决定级的方法更有效的。在这份报纸，我们在场当时，为特征级的多模式的生物测定的信息熔化的概括正规判别式设计(MGCDP)方法，它最大化intra班的关联展示的multiset最小化在班之间的关联。另外，连续MGCDP(S-MGCDP)和策略也是的平行MGCDP(P-MGCDP)求婚了，它能多于二种生物测定的信息熔化，以便完成更好的鉴定效果。实验在MGCDP方法超过的各种各样的生物测定的数据库表演表现了另外的最先进的特征级的信息熔化途径。

简介：Multidisciplinaryfeasiblemethod(MDF)isconventionalmethodtomultidisciplinaryoptimization(MDO)andwell-understoodbyusers.Itreducesthedimensionsofthemultidisciplinaryoptimizationproblembyusingthedesignvariablesasindependentoptimizationvariables.However,ateachiterationoftheconventionaloptimizationprocedure,multidisciplinaryanalysis(MDA)isnumerouslyperformedthatresultsinextremeexpenseandlowoptimizationefficiency.TheintrinsicweaknessofMDFisduetothetimesthatitloopfixed-pointiterationsinMDA,whichdriveustoimproveMDFbybuildinginexpensiveapproximationsassurrogatesforexpensiveMDA.AnsimpleexampleispresentedtodemonstratetheusefulnessoftheimprovedMDF.ResultsshowthatasignificantreductioninthenumberofmultidisciplinaryanalysisrequiredforoptimizationisobtainedascomparedwithoriginalMDFandtheefficiencyofoptimizationisincreased.

简介：Thispaperanalyzesthelock/capturebehaviorofconstantmodulusapproach(CMA)forblindmultiuserdetection.ConditionsunderwhichCMAlocksonthedesiredsignalandwhenitcapturestheinterfererarederived.

简介：Thekeyofthesubspace-basedDirectionOfArrival(DOA)estimationliesintheestimationofsignalsubspacewithhighquality.Inthecaseofuncorrelatedsignalswhilethesignalsaretemporallycorrelated,anovelapproachfortheestimationofDOAinunknowncorrelatednoisefieldsisproposedinthispaper.TheapproachisbasedonthebiorthogonalitybetweenamatrixanditsMoore-Penrosepseudoinverse,andmadenoassumptiononthespatialcovariancematrixofthenoise.Theapproachexploitsthestructuralinformationofasetofspatio-temporalcorrelationmatrices,anditcangivearobustandpreciseestimationofsignalsubspace,soapreciseestimationofDOAisobtained.Itsperformancesareconfirmedbycomputersimulationresults.

简介：Anewalgorithmofnonuniformitycorrectionforinfraredfocalplanearray(IRFPA)isreported,whichisacombinedalgorithmbasedonboththetwo-pointcorrectionandartificialneuralnetworkscorrection.Thecombinedalgorithmiscalibratedbytwo-pointcorrection,andthecalibratedcorrectioncoefficientsareautomaticallymodifiedbyBPalgorithm.Soitisnotonlycalibrated,butalsoreal-timeprocessed.Inadaptivenonuniformitycorrectionalgorithm,thephenomenaghostartifactandtargetfade-outareavoidedbyedgeextraction.Inordertogetintensifiedimage,themodifiedmedianfiltersareadopted.Thesimulateddataindicatestheproposedschemeisaneffectivealgorithm.

简介：一条新奇测试途径为在可编程的门数组(FPGA)是的地里互连资源(红外)proposed.In测试途径，在FPGA的红外的SB(开关框)被利用了测试IRs.Furthermore，在FPGA的可配置的逻辑块(CLB)也被采用了提高开车能力和差错红外的位置能被监视红外决定联系SBs.As结果，红外能与最小的配置patterns.In最大地被扫描实验，内部地发达的FPGA测试

简介：Toimprovetheperformanceoftransmissionbyreducingthenumberoftransmissionandnetworkoverheadofwirelesssingle-hopnetworks,thispaperpresentsahighefficientmultipacketdecodingapproachfornetworkcoding(EMDNC)inwirelessnetworksaccordingtotheideaofencodingpacketswhichcannotbedecodedandarestoredinbufferbyreceivingnodes,thelostpacketscanberecoveredfromtheseencodedpackets.Comparedwiththenetworkcodingwirelessbroadcastingretransmission(NCWBR),EMDNCcanimprovetheefficiencyofdecodingandreducethenumberofretransmissionandtransmissiondelay.SimulationresultsrevealthatEMDNCcaneffectivelyreducethenumberofretransmissionandnetworkoverhead.