PREDIÇÃO DO NÍVEL DE SOMBREAMENTO UTILIZANDO UM MODELO DE REGRESSÃO LOGÍSTICA MULTINOMIAL EM SISTEMAS DE CRIAÇÃO DE BOVINOS DE CORTE

Main Article Content

Rubia Mara de Oliveira SANTOS
Erlandson Ferreira SARAIVA
Thiago Farias ARCE
Ricardo Ribeiro dos SANTOS

Abstract

In the last few years, the consequences of thermal stress on beef cattle have been a subject of interest to researchers, investors and breeders. This is due to the fact that caloric stress has an impact on the decrease of the animals' weight gain and increases in slaughter time and cost of production, leading to the economic loss to the creators/investors. In this paper, we present an electronic platform developed to
capture environmental measures in livestock systems. Using the observed data obtained by the electronic platform we fit fifteen multinomial logistic regression models to predict the level of shading (Sun, Cloudy and Shadow) that a bovine is in. In order to select the best model we consider the model selection criteria AIC and BIC. As a result, we obtained a model with high predictive accuracy.

Article Details

How to Cite
SANTOS, R. M. de O., SARAIVA, E. F., ARCE, T. F., & SANTOS, R. R. dos. (2019). PREDIÇÃO DO NÍVEL DE SOMBREAMENTO UTILIZANDO UM MODELO DE REGRESSÃO LOGÍSTICA MULTINOMIAL EM SISTEMAS DE CRIAÇÃO DE BOVINOS DE CORTE. Brazilian Journal of Biometrics, 37(3), 378–393. https://doi.org/10.28951/rbb.v37i3.406
Section
Articles