资源描述:
《Plot methods in quantitative structure–property studies》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、ChemicalPhysicsLetters393(2004)327–330www.elsevier.com/locate/cplettPlotmethodsinquantitativestructure–propertystudiesa,*,1bLionelloPogliani,JesusV.deJulian-OrtizaDipartimentodiChimica,UniversitadellaCalabria,87030Rende,ItalybUnidaddeInvestigacionenDise~nodeFarmacosyCon
2、ectividadMolecular,DepartmentodeQuımicaFısica,FacultaddeFarmacia,Av.V.A.Estelless/n,46100BurjassotValencia,SpainReceived19May2004;infinalform17June2004AvailableonlineAbstractFivemodelcasesareanalyzedtounderlinetheimportanceandusefulnessofplotmethodsindetectingeitheranom
3、aliesorinconfirmingthequalityofamodel,which,normally,arenoteasilydetectedbytablesorbystatisticalparameters.PlotmethodsinQSARmodelstudiescanbeasimportantasspectroscopicmethodsinmanyexperimentalstudies.Ó2004ElsevierB.V.Allrightsreserved.1.Introductionationorresidualplot(i)ise
4、asytoread,(ii)iseveneasiertointerpret,(iii)givesanearly‘tactile’checkoftheRecentlytherehasbeenanewawakeninginstatisti-qualityofamodel,and(iv)showswherethemodelfailscalmethodsandanalysismainlycenteredonQSAR/andhowmuchitfails,betterthananystatistics.ItQSPR(QuantitativeStruct
5、ure–Activity/PropertyRela-wouldbeeasierforthereadabilityofanarticleifplotstionships)modelstudies[1–7].Actually,theseconsid-couldbegiventoconfirmandstrengthenthestatisticalerationsonstatisticshaveawiderapplicability,andresults,asplotscan,nearly,beconsideredasthe‘spec-concern
6、manyothertypesofmodelstudies.Acarefultra’ofamodel.Itisapitythatthisnotalwayshappens,readingofallthecitedworksunderlinetheimportanceandthereaderisoftenobligedtogothroughpapersofplotmethodsinsuchkindofstudies,ashasbeenburdenedwithtables,sometimeslongpages,ofexperi-recentlyem
7、phasized[8].Infact,manypapersdisplaythemental,calculated,andcalculatedresidualvalues,pluskeyinformation(i)throughtablesofexperimental–cal-thecorrespondingstatistics.InthisshortLetterwewillculatedseriesofvalues,whicharenoteasytoreadandaddressthisproblemwithsomeexamplestaken
8、from(ii)throughstatisticalparameters,whicharenotalwaysrecentQSAR/QSPRpapers,whereonlyalgo