Estimation of lactation milk yield of Awassi sheep with Artificial Neural Network modeling

We modeled data with Artificial Neural Networks (ANNs) for the prediction of consequent milk yield. Models for ANNs were developed, using back-propagation networks with single-hidden layer and sigmoid activation functions. The input variables of the network were age, number of lactations, monthly milk yield (February-May), and lactation duration. The output variable was lactation milk yield. The modeling results demonstrated excellent agreement between the experimental data and the predicted values, with a high determination coefficient (R2 = 0.9998). Hence, the developed model was able to analyze non-linear multi-variant data with a remarkable performance, using limited number of parameters and a short calculation time. The current model may prove to be an alternative method to control expiration date of milk yield shown on labeling and provide safer food supply to consumers. The major use of this predictive process can be formulation of accurate selection decisions based on prior knowledge of the outcomes.