Body Weight Prediction using Recursive Partitioning and Regression Trees (RPART) Model in Indian Black Bengal Goat Breed: A Machine Learning Approach

Abstract

Background: Live body weight (BW) of livestock animals is truly mirror image of all activities of genetics, nutrition, production, reproduction and health status. Thus, the knowledge of calculating BW is of great importance to the producer and critical for goat farming and business. However, there is an unavailability of suitable scales, leading to inaccuracies in decision-making. The present work aimed to predict the live BW of Indian Black Bengal goat using certain morphometric data. Methods: The live BW and eight body measurement data from 1427 disease free, non-pregnant goats aged 25.87±10.47 months with 2.78±1.21 number of parity were collected. The data were first subjected to stepwise regression analysis to achieve the best-fitted model for BW prediction by comparing coefficient of determination (R2) and determining the combination of body dimensions that explained variation in the dependent variable. Further, Recursive Partitioning and Regression Trees (RPART) model, a machine learning tool was deployed to predict BW using certain body measurements. Result: The results of stepwise regression model clearly indicated that heart girth (HG) and punch girth (PG) measurements influenced live BW mostly, but the predictive capabilities (Low R2) of this statistical model were low. The stepwise regression model could not satisfactorily predict BW due to the problem of multicollinearity. Out of eight independent variables, the most important variables emerged from RPART were only HG and PG based on the largest reduction in overall sums of squares error. RPART generated a decision tree with minimal expected error to precisely predict live BW. Hence, RPART model was found to provide better predictive result than stepwise regression model in accurately predicting BW from body measurement variables in Black Bengal goats.