Probit or Logit? Which is the better model to predict the longevity of seeds?

Página do item simplificado
dc.contributor.authorFaria, Rute Q. de
dc.contributor.authorSantos, Amanda R. P. dos [UNESP]
dc.contributor.authorAmorim, Deoclecio J. [UNESP]
dc.contributor.authorCantao, Renato F.
dc.contributor.authorSilva, Edvaldo A. A. da [UNESP]
dc.contributor.authorSartori, Maria M. P. [UNESP]
dc.contributor.institutionInst Fed Educ Ciencia & Tecnol Goiano
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)
dc.contributor.institutionUniversidade Federal de São Carlos (UFSCar)
dc.date.accessioned2020-12-10T20:04:58Z
dc.date.available2020-12-10T20:04:58Z
dc.date.issued2020-03-01
dc.description.abstractThe prediction of seed longevity (P50) is traditionally performed by the use of the Probit model. However, due to the fact that the survival data are of binary origin (0,1), the fit of the model can be compromised by the non-normality of the residues. Consequently, this leads to prediction losses, despite the data being partially smoothed by Probit and Logit models. A possibility to reduce the effect of non-normality of the data would be to apply the principles of the central limit theorem, which states that non-normal residues tend to be normal as thensample is increased. The Logit and Probit models differ in their normal and logistic distribution. Therefore, we developed a new estimation procedure by using a small increase of thensample and tested it in the Probit and Logit functions to improve the prediction of P50. The results showed that the calculation of P50 by increasing thensamples from 4 to 6 replicates improved the index of correctness of the prediction. The Logit model presented better performance when compared with the Probit model, indicating that the estimation of P50 is more adequate when the adjustment of the data is performed by the Logit function.en
dc.description.affiliationInst Fed Educ Ciencia & Tecnol Goiano, Dept Agr Engn, Campus Urutai,Km 2,5, BR-75790000 Urutai, Go, Brazil
dc.description.affiliationUniv Estadual Paulista, Sch Agr, Dept Prod & Plant Breeding, UNESP, Botucatu Av Univ 3780, BR-18610034 Botucatu, SP, Brazil
dc.description.affiliationUniv Fed Sao Carlos, Campus Sorocaba UFSCar, BR-18052780 Sorocaba, SP, Brazil
dc.description.affiliationUnespUniv Estadual Paulista, Sch Agr, Dept Prod & Plant Breeding, UNESP, Botucatu Av Univ 3780, BR-18610034 Botucatu, SP, Brazil
dc.description.sponsorshipFundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)
dc.description.sponsorshipIdFAPESP: 2016/13126-0
dc.format.extent49-58
dc.identifierhttp://dx.doi.org/10.1017/S0960258520000136
dc.identifier.citationSeed Science Research. Cambridge: Cambridge Univ Press, v. 30, n. 1, p. 49-58, 2020.
dc.identifier.doi10.1017/S0960258520000136
dc.identifier.issn0960-2585
dc.identifier.urihttp://hdl.handle.net/11449/197061
dc.identifier.wosWOS:000547265500008
dc.language.isoeng
dc.publisherCambridge Univ Press
dc.relation.ispartofSeed Science Research
dc.sourceWeb of Science
dc.subjectcentral limit theorem
dc.subjectlink functions
dc.subjectresidual normality
dc.subjectseed conservation
dc.subjectseed viability
dc.subjectsoybean
dc.titleProbit or Logit? Which is the better model to predict the longevity of seeds?en
dc.typeArtigo
dcterms.licensehttp://journals.cambridge.org/action/displaySpecialPage?pageId=4676
dcterms.rightsHolderCambridge Univ Press
unesp.departmentProdução e Melhoramento Vegetal - FCApt

Arquivos

Coleções

Botucatu - FCA - Faculdade de Ciências Agronômicas