Abbas Masoudi; Arash Azarfar
Volume 18, Issue 4 , December 2016, , Pages 877-888
Abstract
In the current study, growth parameters of broiler chickens fed with rice hull were estimated and their final body weigh was predicted using non-linear, spline regression and neural networks models. The experimental treatments were control and dietary inclusion of rice hull at the levels of 2.5, 5 and ...
Read More
In the current study, growth parameters of broiler chickens fed with rice hull were estimated and their final body weigh was predicted using non-linear, spline regression and neural networks models. The experimental treatments were control and dietary inclusion of rice hull at the levels of 2.5, 5 and 7.5 percent. Predicted final body weight estimated by non-linear regression models in the current study was higher in control chicks compare with those fed rice hull containing diets (P<0.05), but similar among the other birds. Inflection point of growth curve occurred earlier in in control chicks than those fed hull rice containing diets (P<0.05), but increasing hull rice in the diet level had no effect on this parameter. The highest and lowest body weight at inflection point observed in birds fed control and those fed diet containing 5 percent of rice hull, respectively (P<0.05). Feeding chicks with diets containing rice hull decreased parameters b of spline regression model compared with control birds, while parameter c was only lower in birds fed diet containing 7.5 percent hull rice compared with birds on control diet (P<0.05). According to our results, spline regression model is more efficient than the non-linear and artificial neural network models to predict body weight of broiler chicks fed with diets containing rice hull at day 42 of age.