Predication of Dairy Milk Production Using Machine Learning Techniques

Abstract

This paper proposes an automated model based on the machine learning (ML) technique to predict cows’ dairy milk production. For predicting milk production, the factors which are considered are the health condition (HC) of cows, feed intake capacity (FIC), and expected relative milk yield (ERMY). Based on the deviations between the observed and the average values, the cow’s health condition is determined. The Artificial Butterfly Optimization (ABO) algorithm is used to estimate the Woods parameters. The objective function of ABO minimizes the root mean squared error (RMSE) of the average daily milk yield for each farm. Finally, artificial neural network (ANN) model was applied based on the variables: HC, FIC, and ERMY and the other parameters like age at calving, the month of calving, the days in milk after calving, and the lactation number. The experimental results show that the proposed ANN-ABO algorithm attains the highest accuracy than the ANN-GA, ANN, and SVM algorithms.