Webbased interactive 2D3D medical image processing and visualization software Land Use Policy.docx

Webbased interactive 2D3D medical image processing and visualization software Land Use Policy.docx

《Webbased interactive 2D3D medical image processing and visualization software Land Use Policy.docx》由会员分享，可在线阅读，更多相关《Webbased interactive 2D3D medical image processing and visualization software Land Use Policy.docx（37页珍藏版）》请在冰豆网上搜索。

Webbasedinteractive2D3DmedicalimageprocessingandvisualizationsoftwareLandUsePolicy

Cleanerproductionapplicationasasustainableproductionstrategy,inaTurkishPrintedCircuitBoardPlant  OriginalResearchArticle

Resources,ConservationandRecycling,Volume54,Issue10,August2010,Pages744-751

BaşakBüyükbay,NilgunCiliz,GunEvrenGoren,AydinMammadov

 Closepreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

AbstractAbstract|Figures/TablesFigures/Tables|ReferencesReferences

Abstract

Duringthelastseveraldecades,theamountofwastefromelectronicandelectricalequipmentmanufacturinghasgrownsubstantiallyandasaresult,theenvironmentalconcernsaboutthisindustryhavealsobeenincreasing.Asoneofthemainbranchesoftheelectronicsindustry,PrintedCircuitBoards（PCB）manufactoryconsumeslargeamountsofhazardouschemicalsandprocessrinsingwaterthatcreatesseverepollutionloadinprocesswastewater.Withinthiscontext,thecleanerproduction（CP）applicationthroughresourceconservationandwastereductionatsourcehasbeenappliedfortheselectedPCBproductionplant.TheprioritizedCPoptionsincluded（a）plasmadesmearasatechnologymodificationthateliminatesbothwaterandhazardouschemicalconsumption,（b）leadfreecoatingsasarawmaterialsubstitution/conservationoption,（c）ammoniacaletchantrecovery,（d）microetchantreuseand（e）drag-outrecoverybydrainboardapplicationason-siterecycle/reuse/recoveryoptions.Thetechnical,environmentalandeconomicalevaluationofCPoptionsindicatedthatrecommendedCPoptionsarefeasibleinvestmentsfortheproductionprocessesthatleadtonotonlycopperrecoverybutalsominimumetchantchemical,electroplatingmetalandrinsingwaterconsumption.TheresultsalsoindicatedthattheimplementationoftherecommendedCPoptionsprovidedefinitedecreaseintheend-of-pipetreatment（eop）cost.Thefeasibilityanalyseswerecarriedoutbasedontheinternalrateofreturn（IRR）valuesandpaybackperiodsoftheselectedCPoptionsthatweregroupedinto5-yearmiddleterm（optionshaving1.5–2yearpaybackperiodand35–82%ofIRRvalue）and7-yearlongterm（optionhavinga8yearandabovepaybackperiodand−13%ofIRRvalue）actionplansstartingfromtheyear2007.TwooftheproposedCPoptions,leadfreecoatingsanddrag-outrecoverybydrainboardapplication,havebeensuccessfullyimplementedbytheselectedPCBplantsincetheyear2007.

ArticleOutline

1.Introduction

2.IntegrationofcleanerproductionassessmentintotheselectedPCBproductionprocesses

3.AssessmentphaseandevaluationofCPoptions

4.Feasibilityresultsoftheoptions

4.1.Substitutionofpermanganatedesmearwithplasmadesmear

4.2.Leadfreesolderforpatternplatingandhotairsolderleveling（HASL）

4.3.Alkalineetchantrecoveryprocess

4.4.Thecontinuous-flowsystemformicroetchantreuseandcopperrecovery

4.5.Drainboardapplicationfordrag-outrecoveryinpatternplating

5.Discussionandconclusions

Acknowledgements

References

Purchase

$31.50

377

Ruleextractionfromsupportvectormachines:

Areview  OriginalResearchArticle

Neurocomputing,Volume74,Issues1-3,December2010,Pages178-190

NahlaBarakat,AndrewP.Bradley

 Closepreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

AbstractAbstract|Figures/TablesFigures/Tables|ReferencesReferences

Abstract

Overthelastdecade,supportvectormachineclassifiers（SVMs）havedemonstratedsuperiorgeneralizationperformancetomanyotherclassificationtechniquesinavarietyofapplicationareas.However,SVMshaveaninabilitytoprovideanexplanation,orcomprehensiblejustification,forthesolutionstheyreach.Ithasbeenshownthatthe‘black-box’natureoftechniqueslikeartificialneuralnetworks（ANNs）isoneofthemainobstaclesimpedingtheirpracticalapplication.Therefore,techniquesforruleextractionfromANNs,andrecentlyfromSVMs,wereintroducedtoamelioratethisproblemandaidintheexplanationoftheirclassificationdecisions.Inthispaper,weconductaformalreviewoftheareaofruleextractionfromSVMs.Thereviewprovidesahistoricalperspectiveforthisareaofresearchandconceptuallygroupsandanalyzesthevarioustechniques.Inparticular,weproposetwoalternativegroupings;thefirstisbasedontheSVM（model）componentsutilizedforruleextraction,whilethesecondisbasedontheruleextractionapproach.Theaimistoprovideabetterunderstandingofthetopicinadditiontosummarizingthemainfeaturesofindividualalgorithms.Theanalysisisthenfollowedbyacomparativeevaluationofthealgorithms’salientfeaturesandrelativeperformanceasmeasuredbyanumberofmetrics.Itisconcludedthatthereisnoonealgorithmthatcanbefavoredingeneral.However,methodsthatarekernelindependent,producethemostcomprehensiblerulesetandhavethehighestfidelitytotheSVMshouldbepreferred.Inaddition,aspecificmethodcanbepreferredifthecontextoftherequirementsofaspecificapplication,sothatappropriatetradeoffsmaybemade.ThepaperconcludesbyhighlightingpotentialresearchdirectionssuchastheneedforruleextractionmethodsinthecaseofSVMincrementalandactivelearningandotherapplicationdomains,wherespecialtypesofSVMsareutilized.

ArticleOutline

1.Introduction

1.1.RuleextractionfromSVMs:

themotivation

1.2.RuleextractionfromSVMs:

theproblem

1.3.Overview

2.Supportvectormachineclassifiers:

anoverview

2.1.SoftmarginlinearSVMs

2.2.Non-linearSVMs

3.Ruleextractionfromsupportvectormachines

3.1.Supportvectormachinestraining

3.2.Ruleextraction

3.2.1.MethodsutilizingtheSVMmodelasaclosed-box

3.2.2.MethodsutilizingtheSVsonly

3.2.3.MethodsutilizingSVMSVsandtheseparatinghyper-plane

3.2.3.MethodsutilizingSVs,trainingdataandseparatinghyper-plane

4.Relatedwork:

ruleextractionfromANNstaxonomy

4.1.Translucency

4.2.Quality

4.3.Expressivepower

4.4.Portability

4.5.Algorithmiccomplexity

5.SimilaritiesbetweenANNsandSVMs

6.Discussion

6.1.Region-basedrules

6.2.Decisiontree-basedrules

6.3.Fuzzyruleextraction

6.4.Sequentialcoveringruleextraction

7.Futureresearchquestions

8.Summaryandconclusions

Acknowledgements

AppendixA.Abstractrepresentationfordifferentphasesinruleextractionalgorithms

A.1.Decisiontree-basedruleextraction

A.2.Sequentialcoveringruleextraction

A.3.Fuzzyruleextraction

A.4.Region-basedruleextrtraction

A.4.1.RulExSVMphases

A.4.2.SVM+prototypephases

A.4.3.Hyper-rectangleruleextraction（HRE）phases

References

Vitae

Purchase

$31.50

378

TheInfluenceofProgrammaticChangeonRadiationTherapistResearchCapacity—ASingle-centerCaseStudy  OriginalResearchArticle

JournalofMedicalImagingandRadiationSciences,Volume40,Issue4,December2009,Pages170-177

TaraRosewall,ValerieKelly,JaneHiggins,ShaoHuiHuang,JingYan,JulieWenz,MichaelMilosevic

 Closepreview  |  Relatedarticles  |  Relatedreferenceworkarticles    

AbstractAbstract|Figures/TablesFigures/Tables|ReferencesReferences

Abstract

Purpose:

ToidentifyvariousprogrammaticchangesimplementedinaCanadianradiotherapydepartmenttobuildtherapistresearchcapacity,andtodeterminetheircombinedimpactonquantitativemetricsoftherapistresearchoutput.

Methodsandmaterials:

Thiswasasingle-centercasestudydesign.Programmaticchangeswereretrospectivelyidentifiedfromvariousdepartmentaldocumentarysources.Thosechanges,whichwereactivebetweenJanuary2004andDecember2008andwereimplementedwiththeintentionofincreasingtherapistresearchoutput,werecategorizedbyprimarypurposeaccordingtopublishedcriteriafromtheAlliedHealthProfessionsResearchandDevelopmentActionPlan.Therapistresearchoutputwascollectedoverthesametimeperiodbyanannualdepartment-widee-mailrequestforinformationandverifiedthroughvariousindependentsources.

Results:

Fiveeducationalinitiativeshadthepotentialtobuildtherapistresearchknowledgeandskills（e.g.,journalclub）.Changesimplementedtoprovideinfrastructuretosustaintherapistresearchincludedthecreationofrolesincorporatingaformalresearchcomponent.Fourinitiativeshadthepotentialtopromoteresearchdisseminationandnetworking（e.g.,writinggroup）.Thenumberoftherapistprincipalauthorsincreasedduringthe5years（from4to14perannum）,withapproximately60%ofarticlespublishedininternationalradiationmedicinejournals.Thenumberoftherapistspresentingatconferencesincreasedfrom32in2004to63in2008,with94%ofsubmittedabstractsacceptedforpresentationsin2008.Therapistsaccumulatedover$52,000inpeer-reviewedgrantfundsasprincipalinvestigatorsandtheproportionofresearch-basedtherapistacademicappointmentshasincreasedfrom10%to33%ofappointees.

Conclusion:

InvaluableprogresshasbeenmadeinaCanadianradiotherapydepartmentbycombiningmultipleresearchcapacitybuildingprogrammaticchangestoestablishaculturethatencouragesandsupportstherapistresearchpursuits.Thishasincreasedboththequantityandqualityoftherapistresearchactivity.

Résumé

But:

Recenserlesdifférentschangementsdeprogrammationmisenplacedansunservicederadiothérapiecanadiendanslebutderenforcerlacapacitéderecherchedesthérapeutesetdedéterminerleseffetsdeceschangementssurlesrésultatsdesmesuresquantitativesderecherches.

Méthodesetmatériel:

Ils'agitd'uneétudedecasdansunseulendroit.Onaeffectuéunrelevérétrospectifdeschangementsdeprogram