Data-driven prediction of cattle weight gain for evaluating key growth factors with machine learning approaches

Accurate prediction of cattle weight gain is critical for optimizing herd management, improving productivity, and promoting sustainable livestock practices. Traditional monitoring methods, such as manual data collection and periodic surveys, often lack the precision and timeliness required for reliable forecasting. This study evaluated multiple machine learning models Linear Regression, Decision Tree Regression, XGBoost, and Support Vector Regression (SVM) to predict cattle weight at 36 months and identify the key factors influencing growth. A comprehensive historical dataset was preprocessed to handle missing values, correct inconsistencies, and engineer relevant features. Models were assessed using R-squared (R ² ), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE) to evaluate predictive accuracy and robustness. Results indicated that weights at 30, 24, and 18 months were the strongest predictors of final weight, demonstrating the cumulative effect of early growth. Additional influential factors included breed (notably Zebu), grazing type, and cattle movement frequency, while birth weight showed a negative association, reflecting compensatory growth in lighter-born calves. Environmental variables such as temperature and seasonal conditions had moderate but consistent effects. The Decision Tree model, with the highest predictive performance (R ²  = 0.927; MAE = 2.21 kg; RMSE = 2.95 kg), provided the most interpretable and actionable insights. Linear Regression, XGBoost, and SVM offered complementary predictions but with slightly lower accuracy or interpretability. These findings provide quantitative insights into cattle growth dynamics and demonstrate the utility of machine learning for data-driven decision-making in livestock management. The study supports the adoption of predictive models, improved data collection protocols, and targeted management interventions during critical growth periods to optimize cattle production sustainably.