Comparative study of multiple linear regression and artificial neural network for prediction of first lactation 305-days milk yield in Tharparkar cattle

Abstract

The present investigation was undertaken on 3266 weekly test day milk yield records of first lactation Tharparkar cows spread over a period of 8 years (2012-2020) maintained at Livestock Research Station, Beechwal, Bikaner. The weekly test day milk yields (WTD) were used to develop best multiple linear regressions (MLR) and artificial neural network (ANN) model for prediction of first lactation 305-days milk yield (FL305DMY). Further, the comparison was made between MLR and ANN model based on coefficient of determination (R²) and root mean square error (RMSE). Artificial Neural Network was trained using back propagation algorithms viz. Scaled conjugate gradient (SCG). It has been observed that the coefficient of determination of the models was increased with the addition of test day milk yields as input variables. It was inferred from the study that artificial neural network was better than the multiple linear regression to predict FL305DMY with more than 70% accuracy by almost all the input sets at early as 117^th day of the lactation with lesser value of RMSE in comparison to MLR. Therefore, it is concluded that ANN is a potential tool for the prediction of the first lactation305-days milk yield in Tharparkar cattle than multiple linear regression.