Data-Driven Prediction of Cattle Weight Gain for Evaluating Key Growth Factors with Machine Learning Approaches

Accurate prediction of cattle growth is critical for improving herd management, enhancing productivity, and promoting sustainable agricultural practices. Traditional monitoring methods, including manual data collection and periodic surveys, often lack the precision and immediacy necessary for reliable forecasting. This study addresses these limitations by employing machine learning techniques specifically Linear Regression and Decision Tree Regression to identify and analyse the key factors influencing cattle weight gain. Utilizing a comprehensive historical dataset, data preprocessing steps were undertaken to handle missing values, correct inconsistencies, and engineer relevant features. The models were trained and evaluated using standard metrics such as R-squared (R²), Mean Squared Error (MSE), and Mean Absolute Error (MAE) to ensure robustness and accuracy. Findings indicate that weights recorded at 30, 24, and 18 months are the most significant predictors of final weight at 36 months. Additionally, breed (notably Zebu), grazing methods, and cattle movement frequency demonstrated substantial influence on weight outcomes. Contrarily, birth weight exhibited a negative correlation with final weight, suggesting compensatory growth phenomena in calves with lower birth weights. Environmental variables, including temperature and seasonal conditions, showed moderate but consistent effects. These results provide valuable quantitative insights into cattle growth dynamics and reinforce the potential of machine learning as a tool for data-driven decision-making in livestock management. The study advocates for broader adoption of such predictive models, enhanced data collection protocols, and targeted management interventions during key growth periods to optimize cattle production sustainably.