Artificial Neural Network versus Multiple Regression Analysis for Prediction of Lifetime Milk Production in Sahiwal Cattle

Abstract

Gandhi, R.S., Raja, T.V., Ruhil, A.P. and Kumar, A. 2010. Artificial neural network versus multiple regression analysis for prediction of lifetime milk production in Sahiwal cattle. J. Appl. Anim. Res., 38: 233–237.

First lactation records (1493) of Sahiwal cows comprising of age at first calving, first lactation 305-d or less milk yield, first lactation length, first service period and first dry period were analyzed to predict the lifetime milk production using multiple regression analysis (MRA) and artificial neural network (ANN). The data was divided into three sets i.e. cows retained in the herd up to first three parities (set I), 4^th to 6^th parities (set II) and more than 6^th parities (set III). The accuracy of prediction of lifetime milk production in MRA was lower than the accuracy of ANN in all the test data sets. Also, the root mean square errors of prediction were lower from ANN in comparison to MRA. Higher estimates of accuracy of prediction of lifetime milk yield from ANN in comparison to MRA in both the data sets revealed that this methodology can be used as an alternate approach to predict lifetime milk production.

Key words:

• Artificial neural network
• first lactation traits
• lifetime milk yield
• multiple regression analysis
• Sahiwal cattle