Prediction of First Lactation 305-day Milk Yield Based on Monthly Test Day Records Using Artificial Neural Networks in Sahiwal Cattle

Abstract

In the present study, first lactation 305-day lactation milk yield (FL305DMY) was predicted by artificial neural network (ANN) using monthly test day milk yields records of 588 Sahiwal cows. A total of five monthly test day milk yields (2, 3, 5, 7, 8 monthly test day record) were used in neural networks to train data using Bayesian regularization (BR) algorithm. Results showed that the accuracy of prediction of all the models increased with the addition of test day milk yields as input variables. The best neural network model was able to predict FL305DMY with 93.18% accuracy. Further, comparison was made between multiple linear regression (MLR) and ANN for accuracy of prediction and there was no significant different found between ANN and MLR for prediction of FL305DMY in Sahiwal cows.