download0 view802
twitter facebook

공공누리This item is licensed Korea Open Government License

dc.contributor.author
정승현
dc.contributor.author
정민중
dc.date.accessioned
2019-08-28T07:41:12Z
dc.date.available
2019-08-28T07:41:12Z
dc.date.issued
2012-05-01
dc.identifier.issn
2234-7593
dc.identifier.uri
https://repository.kisti.re.kr/handle/10580/14080
dc.identifier.uri
http://www.ndsl.kr/ndsl/search/detail/article/articleSearchResultDetail.do?cn=JAKO201217355624195
dc.description.abstract
Inthispaper,weproposetouseasupportvectormachine(SVM)fortheclassificationofdesigndata.AlthoughtheSVMisaverypopulartechniqueindatamining,itisrarelyappliedtoanindustrialdesignprocessthatmayrequireinformationregardingthefeasibilityofthedesignpointofinterest.Tocheckthefeasibility,thedesignermustconductexperimentsorcomputersimulations,whichmayincurconsiderablecost.Therefore,theSVMcanbeaneffectivetoolforpredictingfeasibleandinfeasibleregionsbecauseitonlyusesthecumulativedesigndata.Inthispaper,weusedtheSVMtoclassifysampledatasetsdrawnfrommathematicaltestproblemsandfromanair-conditionerpipedesignexample.OurresultsindicatethattheSVMiscapableofveryaccuratelyidentifyingfeasibleandinfeasibleregionsinthedesignspace.Further,
wewereabletoreducethetrainingtimeoftheSVMbyusingthek-meansclusteringalgorithmtoreducetheamountoftrainingdata,takingadvantageofthepowerfulgeneralizationabilitiesoftheSVM.Consequently,weconcludethattheSVM
canbeaneffectivetooltoassessfeasibilityatcertaindesignpoints,avoidingsomeofthehighcomputationalcostsoftheanalysis.
dc.language
eng
dc.relation.ispartofseries
Internationaljournalofprecisionengineeringandmanufacturing
dc.title
Feasibility Classification of New Design Points Using Support Vector Machine Trained by Reduced Dataset
dc.citation.number
5
dc.citation.startPage
739
dc.citation.volume
13
dc.subject.keyword
Supportvectormachine(SVM)
dc.subject.keyword
Feasibilityclassification
dc.subject.keyword
K-meansclustering
dc.subject.keyword
Air-conditionerpipedesignproblem
Appears in Collections:
7. KISTI 연구성과 > 학술지 발표논문
Files in This Item:
There are no files associated with this item.

Browse