Data analysis is pervasive throughout business, engineering and science. Very often the data to be analyzed is nonnegative, and it is often preferable to take this constraint into account in the analysis process. Here we are concerned with the application
NonnegativeMatrixFactorizationfor
SpectralDataAnalysis
V.PaulPauca
J.Piper
RobertJ.Plemmons
Abstract
Dataanalysisispervasivethroughoutbusiness,engineeringandscience.Veryoftenthedatatobeanalyzedisnonnegative,anditisoftenpreferabletotakethisconstraintintoaccountintheanalysisprocess.Hereweareconcernedwiththeapplicationofanalyzingdataobtainedusingastronomicalspectrometers,whichprovidespectraldatawhichisinherentlynonnegative.Theidenti cationandclassi cationofspaceobjectsthatcannotbeimagedinthenormalwaywithtelescopesisanimportantbutdi cultproblemfortrackingthousandsofobjects,includingsatellites,rocketbodies,debris,andasteroids,inorbitaroundtheearth.Inthispaperwedevelopane ectivenonnegativematrixfactorizationalgorithmwithnovelsmoothnessconstraintsforunmixingspectralre ectancedataforspaceobjectidenti cationandclassi cationpurposes.Promisingnumericalresultsarepresentedusinglaboratoryandsimulateddatasets.
Keywords:Nonnegativematrixfactorization,spectraldata,blindsourceseparation,datamining,spaceobjectidenti cationandclassi cation.
1Introduction
Weareconcernedwithmethodsbasedonnonnegativematrixfactorization(NMF)fortheanalysisofspectralre ectancedataassociatedwithspaceobjects.Aswillbedescribedlater,
DepartmentofComputerScience,WakeForestUniversity,Winston-Salem,NC27109.HisresearchwassupportedinpartbytheAirForceO ceofScienti cResearchundergrantFA49620-03-1-0215,andbytheArmyResearchO ceundergrantDAAD19-00-1-0540.Email:paucavp@wfu.edu
DepartmentofComputerScience,WakeForestUniversity,Winston-Salem,NC27109.HisresearchwassupportedinpartbytheAirForceO ceofScienti cResearchundergrantF49620-02-1-0107.Email:pipejw02@wfu.edu
DepartmentsofComputerScienceandMathematics,WakeForestUniversity,Winston-Salem,NC27109.HisresearchwassupportedinpartbytheAirForceO ceofScienti cResearchundergrantF49620-02-1-0107andbytheArmyResearchO ceundergrantDAAD19-00-1-0540.Email:plemmons@wfu.edu
1
Nonnegative matrix factorization and applications_专业资料。Data analysis is ...2.2 Image and Spectral Data Processing Digital images are represented as ...
Indexing terms: Nonnegative Matrix Factorization (NMF), Blind Source Separation (BSS), Data Mining and Analysis Introduction: NMF and its extended version, ...
Keywords nonnegative matrix factorization, mixture model, sparseness, gene expression data analysis, pattern classification Genome research has become a very ...
Nonnegative Matrix Factorization with Earth Mover's Distance Metric for Image Analysis Sandler, R. Lindenbaum, M. Page(s): 1590 - 1602 Digital Object ...
Plemmons, Nonnegative matrix factorization for spectral data analysis, Linear Algebra and its Applications, vol. 416 (2006), pp. 2947, 2006. P.O. Hoyer...
Nonnegative Matrix Factorization (NMF) has been widely used in machine learning and data mining....Z (V, V ) = ij 3.3 Convergence Analysis ?...
Detect and Track Latent Factors with Online Nonnegative Matrix Factorization Detecting and tracking latent factors from temporal data is an important task. ...
[cs.NA] 27 Sep 2007 February 1, 2008 Abstract Nonnegative matrix factorization (NMF) has become a prominent technique for the analysis of image databases...
or more voices by nonnegative matrix factorization...[16], music analysis [17] and sound ...spectral peaks as if they represented a ...
The spectral norm of a n... 12页 免费喜欢... pattern analysis, clustering, and cancer class ... 1 Introduction Nonnegative Matrix Factorization (...
热门文档