Graded Buffalypso male growing curve estimation using non lineal mixed models

Abstract Text: A database of 1188 observations with 11 monthly weightings from weaning to 18 month of age of 108 Buffalypso graded males buffaloes from “Los Naranjos” enterprise at Artemisa province was utilized. A non-lineal mixed model (PROC NLMIXED from SAS, 2007) was applied where fixed effects were: weaning month and year. Individual weights were adjusted to the corresponding ages Different non lineal models variants are studied: Logistic, Gompertz, Brody and Von Bertanlaffy including to β0 parameter the random component b0. The β0, β1 and β2 parameters were considered as fixed effects. The goodness of fit criteria were: maximum likelihood criterion, Akaike information criterion and bayesian information criterion. The Logistic and Gompertz models attained the best adjustments to describe growing curves of the population under study.

Keywords: buffaloes growing curve