简介：

简介：

简介：Themotivationofdataminingishowtoextracteffectiveinformationfromhugedatainverylargedatabase.However,someredundantandirrelevantattributes,whichresultinlowperformanceandhighcomputingcomplexity,areincludedintheverylargedatabaseingeneral.So,FeatureSubsetSelection(FSS)becomesoneimportantissueinthefieldofdatamining.Inthisletter,anFSSmodelbasedonthefilterapproachisbuilt,whichusesthesimulatedannealinggeneticalgorithm.Experimentalresultsshowthatconvergenceandstabilityofthisalgorithmareadequatelyachieved.

简介：ＡＮＯＶＥＬＤＥＦＩＮＩＴＩＯＮＯＦＦＯＲＭＦＥＡＴＵＲＥＧａｏＦｅｉ；ＹｅＳｈａｎｇｈｕｉＡＮＯＶＥＬＤＥＦＩＮＩＴＩＯＮＯＦＦＯＲＭＦＥＡＴＵＲＥ￥ＧａｏＦｅｉ；ＹｅＳｈａｎｇｈｕｉＡｂｓｔｒａｃｔＢｅｉｎｇａｎｅｓｓｅｎｔｉａｌｃｏｎｃｅｐｔｉ...

简介：

简介：Inthisnotewecharacterizethegeometricfactureofa(μ,r,k)-FES.Namely,foraC~μtriangularin-terpolationschcmewithC~rvertexdata,anyangleofthemacrotrianglemustbedividedintoatleast(μ+1)/(r+1-μ)parts.

简介：Inapplicationsoflearningfromexamplestoreal-worldtasks,featuresubsetselectionisimportanttospeeduptrainingandtoimprovegeneralizationperformance.Ideally,aninductivealgorithmshouldusesubsetoffeaturesassmallaspossible.Inthispaperhowever,theauthorsshowthattheproblemofselectingtheminimumsubsetoffeaturesisNP-hard.Thepaperthenpresentsagreedyalgorithmforreaturesubsetselection.Theresultofrunningthegreedyalgorithmonhand-writtennumeralrecognitionproblemisalsogiven.

简介：Thispaperpresentsanovelapproachtofeaturesubsetselectionusinggeneticalgorithms.Thisapproachhastheabilitytoaccommodatemultiplecriteriasuchastheaccuracyandcostofclassificationintotheprocessoffeatureselectionandfindstheeffectivefeaturesubsetfortextureclassification.Onthebasisoftheeffectivefeaturesubsetselected,amethodisdescribedtoextracttheobjectswhicharehigherthantheirsurroundings,suchastreesorforest,inthecoloraerialimages.Themethodologypresentedinthispaperisillustratedbyitsapplicationtotheproblemoftreesextractionfromaerialimages.

简介：处理技术的计算机图象被用来在木头表面上完成缺点图象的特征抽取。由缺点的灰色的价值的计算。有用鉴别缺点被完成了的三个特征数据。实验显示那这样对木头表面上的缺点的自动化识别有效。

简介：Theanalysisoftheradiatednoiseofvesselsgiveninthispapershowssomestrongsuperposedlinecomponentsinlowfrequencyspeetrumbelow100Hzoccurringatdiscretefrequencieswhichcorrespondwiththerotationspeedofpropellershaft,orpropellerbladefrequency,ortheirharmonicfre-quencies.sincethelinecomponentsreflectpropeller＇workingcharacteristics,thepropller＇sfeaturescanbeextracteddirectlyfromlow-frequencylinecom-ponentsinadditiontodemodulatedlinecomponent.Sotherearetwowaystoextractthefeatures,oneisdirectway,theotherisdemodulationway.Detec-tionperformanceofthelinecomponentinbackground-noiseisdiscussedinthispaper.ThesignallevelisdefinedasthedifrerencebetweenthePDF＇s(ProbabilityDensityFunction)meanofthepeakofthelinecomponentandPDF＇smeanorthebackground-noise.Indircetwaythesignallevelofthelinecomponentisproportionaltothesignalnoiseratio(S/N).Indemodulationwaythesignallevelofdemodulatedlinecomponent

简介：IntroductionDuringthelasttwodecadestherehavebeenremarkabledevelopmentsinEnglishlanguageteachinginChina.Withanever-increasingemphasisbeingputonthecommunicativecompetenceofthelearners,variousmethodsandtechniqueshavebeentriedtoimprovethestudents’abilitytocommunicateinEnglish.ThisarticlelooksatthevalueofusingWesternfeaturefilmstohelpstudentsdevelopthiscommunicativecompetence.

简介：

简介：WegiveanintroductionforthebackgroundandmotivationoftheIntegratedPhotonics:ChallengesandPerspectivesfeature.Averybriefsummaryforthefiveinvitedreviewarticlescollectedinthisfeatureissueisalsogiven.

简介：特征定义的适应性是特征技术在设计过程成功应用的关键和核心问题．本文基于提出了面向设计过程的特征定义层次结构及特征定义方法，将特征定义用应用特征、形状特征和几何表示３个相对独立的层次来表示，并分别利用特征语义解释器和改进的几何关联图实现层次间的有效连接和相互转换，这不仅消弱了特征定义对应用的依赖性而更具广泛性，也为基于特征的并行设计过程模型的建立提供了基础．

简介：Translationinitiationsites(TISs)areimportantsignalsincDNAsequences.InmanypreviousattemptstopredictTISsincDNAsequences,threemajorfactorsaffectthepredictionperformance:thenatureofthecDNAsequencesets,therelevantfeaturesselected,andtheclassificationmethodsused.Inthispaper,weexaminedifferentapproachestoselectandintegraterelevantfeaturesforTISprediction.Thetopselectedsignificantfeaturesincludethefeaturesfromthepositionweightmatrixandthepropensitymatrix,thenumberofnucleotideCinthesequencedownstreamATG,thenumberofdownstreamstopcodons,thenumberofupstreamATGs,andthenumberofsomeaminoacids,suchasaminoacidsAandD.Withthenumericaldatageneratedfromthesefeatures,differentclassificationmethods,includingdecisiontree,naiveBayes,andsupportvectormachine,wereappliedtothreeindependentsequencesets.Theidentifiedsignificantfeatureswerefoundtobebiologicallymeaningful,whiletheexperimentsshowedpromisingresults.

简介：

简介：Traditionalsequenceanalysisdependsonsequencealignment.Inthisstudy,weanalyzedvariousfunctionalregionsofthehumangenomebasedonsequencefeatures,includingwordfrequency,dinucleotiderelativeabundance,andbase-basecorrelation.Weanalyzedthehumanchromosome22andclassifiedtheupstream,exon,intron,downstream,andintergenicregionsbyprincipalcomponentanalysisanddiscriminantanalysisofthesefeatures.Theresultsshowthatwecouldclassifythefunctionalregionsofgenomebasedonsequencefeatureanddiscriminantanalysis.

简介：

简介：Inthispaper,afacialfeatureextractingmethodisproposedtotransformthree-dimension(3D)headimagesofinfantswithdeformationalplagiocephalyforassessmentofasymmetry.Thefeaturesof3Dpointcloudsofaninfant’scraniumcanbeidentifiedbylocalfeatureanalysisandatwo-phasek-meansclassificationalgorithm.The3Dimagesofinfantswithasymmetriccraniumcanthenbealignedtothesamepose.Themirroredheadmodelobtainedfromthesymmetryplaneiscomparedwiththeoriginalmodelforthemeasurementofasymmetry.Numericaldataofthecranialvolumecanbereviewedbyapediatriciantoadjustthetreatmentplan.Thesystemcanalsobeusedtodemonstratethetreatmentprogress.

简介：Alotof3Dshapedescriptorsfor3Dshaperetrievalhavebeenpresentedsofar.Thispaperproposesanewmechanism,whichemploysseveralexistingglobalandlocal3Dshapedescriptorsasinput.Withthesparsetheory,somedescriptorswhichplaythemostimportantroleinmeasuringsimilaritybetweenquerymodelandthemodelinthedatasetareselectedautomaticallyandanaffinitymatrixisconstructed.Spectralclusteringmethodcanbeimplementedtothisaffinitymatrix.Spectralembeddingofthisaffinitymatrixcanbeappliedtoretrieval,whichintegratingalmostalltheadvantagesofselecteddescriptors.Inordertoverifytheperformanceofourapproach,weperformexperimentalcomparisonsonPrincetonShapeBenchmarkdatabase.Testresultsshowthatourmethodisapose-oblivious,efficientandrobustnessmethodforeithercompleteorincompletemodels.