Ajuste de modelos de crescimento para frangos machos

Abstract

Among the animals raised for cutting, the poultry has shown a great rise, with emphasis on the production of chickens. In Brazil, this market moves millions of reais, since, besides being a major consumer, the country is the largest exporter of this protein. One of the most used lineages in the poultry cutting activity is the Cobb 500r line, because it has high rates of profitability. For this reason, understanding the growth of this animal helps in the production and management, and this can be done through growth curves using non-linear models. The objective of this work was to analyze the adjustments of the nonlinear models of Brody, Gompertz, Log´ıstico and von Bertalanffy in the description of the growth of male chickens of Cobb 500r. R statistical software was used for the analyzes. The fit quality of the models was compared using the following evaluators: coefficient of determination (R2 ), Corrected Akaike Information Criterion (AICc), deviation (DP R) and the mean prediction error (MEM). The parameters of the models were significant, according to the t test, at the 5% level. Then the Shapiro-Wilk, Durbin-Watson and Breusch-Pagan tests were performed for residue analysis.It was not possible to estimate the parameters of the Brody model, since there was no convergence of the algorithms, however, the other models fitted well to the data and von Bertalanffy presented better adjustment quality for the chickens weight of the Cobb 500r, as it presented lower values of AICc, DP R and MEM, in addition to a larger R2 .