Comparative Study of Feed-Forward Neuro-Computing with Multiple Linear Regression Model for Milk Yield Prediction in Dairy Cattle

Abstract

The main objective of this work is to compare the accuracy of artificial neural networks (ANNs) and multiple linear regression (MLR) model for prediction of first lactation 305-day milk yield (FL305DMY) using monthly test-day milk yield records of 443 Frieswal cows. We have compared four versions of feed forward algorithm with conventional statistical model. The performancre of ANN is found to be better than the MLR model for milk yield prediction. The Bayesian regularization neural network model was able to predict milk yield with 85.07% accuracy as early as 126th day of lactation. It has been found that R 2 value of the models increases with increase in the number of test-day milk yield records.