Comparison of different machine learning algorithms for weight prediction

In this study, it was aimed to compare the prediction performances of Random Forest (RF), Extreme Gradient Boost (XGBoost), LightGBM, Gradient Boosting Machine (GBM) and Support Vector Machine (SVM) machine learning algorithms for weight estimation in Romanov lambs. For this purpose, Cidago height, Rump height, Body length, Chest depth, Chest depth, Chest girth, Chest girth between the shoulder blades, Chest width between the shoulder blades and live weights of 50 Romanov lambs born from 285 Romanov ewes reared in Niksar district of Tokat province were used. Mean Square Error (MSE), Root Mean Squared Error (RMSE) and coefficient of determination (R²) values were used to compare the prediction performance of the methods. As a result of the study, it was determined that Gradient Boosting Machine (GBM) algorithm had the best performance and accuracy with the lowest MSE (0.72801), RMSE (0.99534) and the highest R² (0.98022) values in the live weight estimation of Romanov lambs. Additionally, Random Forest (RF), Extreme Gradient Boost (XGBoost), LightGBM, and Support Vector Machine (SVM) algorithms also demonstrated high performance and provided effective results in predicting live weight.