Smarter ways to predict rabbit body weight across multiple breeds

Morphometric measurement is essential in the determination of breeding program zones that need to be improved.

This research aims to compare the precision of morphometric measurement to linear models, such as regression analysis and machine learning methods, such as Random Forest (RF), to improve the precision of live weight estimation in animal breeding programs. A total of 228 rabbits were used in the current study, and they comprised the following breeds:39 Satin, 40 Rex, 40 New Zealand White, 29 Hyla, 40 Hycole, and 40 Reza were utilized for the study. Each rabbit was measured on body weight, head (width and length), chest circumference, body length, and hip width. Stepwise regression and linear regression analyses were conducted using the lm function in R version 4.4.1. For the RF algorithm, the caret and randomForest packages were utilized to build and evaluate the model.

In this study, linear regression [R-squared value of 0.82 and an Root Mean Squared error (RMSE) of 300.16] outperformed RF (R-squared value of 0.8 and an RMSE of 326.37) in predicting rabbit body weight based on morphometric measurements. The results showed that chest circumference and body length were the most influential predictors, with the largest coefficients and highest significance levels, and the IncNodePurity illustration showed head length (IncNodePurity: 19388974) emerged as an important factor in predicting body weight.

The Linear regression model showed superior results compared to the RF model in predicting rabbit body weight based on morphometric measurements.