Artificial neural network and non-linear logistic regression models to fit the egg production curve in commercial-type broiler breeders

The aim of the present study was to investigate the possibility of using artificial neural network (ANN) and non-linear logistic regression models to estimate egg production curves in broiler breeders. Data were collected from 8 flocks for 64 weeks in a commercial broiler breeder farm. Non-linear regression models were determined using the procedure of Guess-Newton Iteration, whereas AAN models were developed with the Statistica Neural Networks software. The accuracy of the models was determined by mean square error (MSE), mean absolute deviation, mean absolute percentage error and bias, respectively. The results showed that the estimates of the ANNs were more accurate compared to other non-linear regression models. Lower MSE)model Training and Testing: 0.001 and 0.002, respectively) and the higher R-2 (model Training and Testing: 0.999 and 0.996, respectively) were obtained using the ANN model indicating that this model is the best to describe the process of egg production in broiler breeders. The advantage of using neural networks is that they can be fitted to any kind of dataset and do not need to model assumptions such as those required in the non-linear logistic models. However, it may be more practical to ignore the relevance of parameter estimates and to focus on the ability to predict responses.