Acessibilidade / Reportar erro

Use of regression techniques in the evaluation, in beef cattle, of feed conversion into product: proposition of method and nutritional significance

The objective of this study was to propose and discuss a method to evaluate the efficiency of feed conversion into product from beef cattle based on regression techniques. The mathematical and statistical procedures were developed using a database composed of measurements of dry matter intake and average daily gain of 380 pure or crossbred zebu animals in 15 different experiments in the database from the BR-CORTE nutritional system. Two approaches were presented, based on linear and non-linear regression, with average daily gain and dry matter intake used as independent and dependent variables, respectively. The feed utilization by the animal was stratified into maintenance demand and real efficiency of conversion into product. The maintenance demand presents similar interpretation in both approaches. When linear solution is considered, the real efficiency of product conversion is assumed to be steady and independent on average daily gain. On the other hand, when the non-linear solution is adopted, the real efficiency of product conversion is assumed to be variable and dependent on gain. Alaike information criterion was proposed to decide which model suits best. The evaluation of the efficiency of transformation of feed into product, in beef cattle, by regression techniques allows a broader understanding of system efficiency for it enables stratification of maintenance and production demands. Several inconvenient aspects concerning ratio indexes, such as food conversion ratio, are avoided and the nutritional inferences become more accurate and precise.

Akaike information criterion; food conversion ratio; linear regression; non-linear regression


Sociedade Brasileira de Zootecnia Universidade Federal de Viçosa / Departamento de Zootecnia, 36570-900 Viçosa MG Brazil, Tel.: +55 31 3612-4602, +55 31 3612-4612 - Viçosa - MG - Brazil
E-mail: rbz@sbz.org.br