简介:摘要本文在将数值型数据标准化的基础上,将分类数据细分为二元数据和类型数据,并用相异度系数距离计算分类数据之间的距离,并且赋予二元和类型数据相应的权重,来改进k-prototypes聚类算法,使该算法满足不同要求的混合属性数据聚类,最后通过C#语言,在ArcEngine2010版本上实现。
简介:ToimprovethecurrentGISfunctionsindescribinggeographicobjectswithfuzziness,thispaperbeginswithadiscussiononthedistancemeasureofspatialobjectsbasedonthetheoryofsetsandanintroductionofdilationanderosionoperators.Undertheassumptionthatchangesofattributesinageographicregionaregradual,theanalyticexpressionsforthefuzzyobjectsofpoints,linesandareas,andthedescriptionoftheirformalstructuresarepresented.Theanalyticmodelofgeographicobjectsbymeansoffuzzyfieldsisdeveloped.Wehaveshownthatthe9-intersectionmodelproposedbyEgenhoferandFranzosa(1991)isaspecialcaseofthemodelpresentedinthepaper.
简介:Thispaperpresentsanewfuzzymultiplecriteria(bothqualitativeandquantitative)decision-making(MCDM)methodbasedonfuzzyrelationaldegreeanalysis.Theconceptsoffuzzysettheoryareusedtoconstructaweightedsuitabilitydecisionmatrixtoevaluatetheweightedsuitabilityofdifferentalternativesversusvariouscriteria.Thepositiveidealsolutionandnegativeidealsolutionarethenobtainedbyusingamethodofrankingfuzzynumbers,andthefuzzyrelationaldegreesofdifferentalternativesversuspositiveidealsolutionandnegativeidealsolutionarecalculatedbyusingtheproposedarithmetic.Finally,therelativerelationaldegreesofvariousalternativesversuspositiveidealsolutionarerankedtodeterminethebestalternative.Anumericalexampleisprovidedtoillustratetheproposedmethodattheendofthispaper.