文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 自然科学 > 数学 > NIR and mass spectra classification Bayesian methods for wavelet-based feature selection
侵权投诉

NIR and mass spectra classification Bayesian methods for wavelet-based feature selection

ChemometricsandIntelligentLaboratorySystems77(2005)139–

NIR and mass spectra classification Bayesian methods for wavelet-based feature selection

NIR and mass spectra classification Bayesian methods for wavelet-based feature selection

NIR and mass spectra classification Bayesian methods for wavelet-based feature selection

148

doc.guandang.net

NIRandmassspectraclassification:Bayesianmethodsfor

wavelet-basedfeatureselection

MarinaVannuccia,T,NaijunShab,PhilipJ.Brownc

b

DepartmentofStatistics,TexasA&MUniversity,UnitedStates

DepartmentofMathematicalSciences,UniversityofTexasatElPaso,UnitedStates

c

InstituteofMathematicsandStatistics,UniversityofKent,Canterbury,UK

Receivedinrevisedform10September2004;accepted10October2004

Availableonline4March2005

a

Abstract

Herewefocusonclassificationproblemsthatinvolvefunctionalpredictors,specificallyspectraldata.Oneofourpracticalcontextsinvolvestheclassificationofthreewheatvarietiesbasedon100nearinfra-redabsorbances.Thedatasetconsistsofatotal117samplesofwheatcollectedduringastudythataimedatexploringthepossibilityofusingNIRspectratoassignunknownsamplestothecorrectvariety.Inanotherexamplewelookatserumspectrafrom162ovariancancerand91controlsubjectsgeneratedthroughsurfaceenhancedlaserdesorptionionizationtime-to-flightmassspectrometry(SELDI-TOF).Weemploywavelettransformsasatoolfordimensionreductionandnoiseremoval,reducingspectratowaveletcomponents.WethenuseprobitmodelsandBayesianmethodsthatallowthesimultaneousclassificationofthesamplesaswellastheselectionofthediscriminatingfeaturesofthespectra.Inbothexamplesourmethodisabletofindverysmallsetsoffeaturesthatleadtogoodclassificationresults.D2005ElsevierB.V.Allrightsreserved.

Keywords:Bayesianvariableselection;Discrimination;Multinomialprobitmodels;NIRspectra;Proteomicdata;Wavelettransforms

1.Introduction

Classificationproblemswithfunctionalpredictordatahasbecomeprominentinmanyfieldsinrecentyears.Nearinfrared(NIR)spectroscopyisusedinmanyareas,suchasintheanalysisoffoodanddrinkandofpharmaceuticalproducts.ANIRspectrumofasampleistypicallymeasuredbymodernscanninginstrumentsathundredsofequallyspacedwavelengths.Theinformationinthecurveisusedtopredictthechemicalcompositionofthesamplebyextractingtherelevantinformationfrommanyover-lappingpeaks.Osborneetal.[1]describestandardapproaches,suchaslineardiscriminantanalysis(LDA).Thesemethodscanfailwithmanyvariablesanddifferentapproachesneedtobetaken,seeforexampleKrzanowski

TCorrespondingauthor.

E-mailaddresses:mvannucci@stat.tamu.edu(M.Vannucci)8nsha@utep.edu(N.Sha)8Philip.J.Brown@kent.ac.uk(P.J.Brown).0169-7439/$-seefrontmatterD2005ElsevierB.V.Allrightsreserved.doi:10.1016/j.chemolab.2004.10.009

etal.[2].AcommonsolutionistoreducethedimensionofthepredictormatrixbyusingprincipalcomponentsandthenapplyLDA.

Recentadvancesinfunctionalgenomicshavemadeitpossibletomeasuretheexpressionlevelsofthousandsofgenesorproteins.Proteomictechnologyisnon-invasiveandrequiressmallamountsofbiologicalmaterial(tissueorbloodsamples).Proteomicsisoftenusedforbiomarkerdiscovery,toidentifyproteinslinkedtodiseasestatus,responsetotherapy,orclinicalprognosis.Amassspectrumcanberepresentedasacurvewherethex-axisindicatestheratiooftheweightofaspecificmoleculetoitselectricalcharge(m/z)andthey-axisisthesignalintensityforthesamemoleculeasameasureoftheabundanceofthatmoleculeinthesample.Proteomicspectraarecharacterizedbymanypeaks,mostofwhichcorrespondtoproteinsorproteinfragments(peptides).Theidentificationofpeaksrelatedtoaspecificoutcome,forexamplepeaksthatdiscriminatesamplesorthatpredictaclinicalresponse,is

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email