Predicting the Weight of Livestock Using Machine Learning

In the management of the cattle farming industry, the prediction of livestock weight has an important role involving multidimensional aspects such as industry, health, and economy, as well as its impact on the economy of cattle breeders. Cow weight is one of the central elements that directly influences economic welfare in the livestock sector. By utilizing artificial intelligence models, farmers can optimize their business results, achieve business sustainability, and experience a positive impact on the livestock economy. In this research, eight methods were used, namely Random Forest, Linear Regression, Ridge Regression, Lasso Regressor, K Neighbors Regressor, Decision Tree Regressor, Gradient Boosting Regressor, and AdaBoost Regressor. The results show that the Linear Regression method performs best with a Mean Absolute Error (MAE) value of 17.8 kg and an R-square of 0.83, outperforming the performance of other algorithmic methods. The analysis results also show that predicting cow weight using the Linear Regression method shows a significant relationship between morphometric variables and live weight. The variable chest circumference (CG), with a correlation coefficient of 0.87, width in maclock (WIM) of 0.82, and chest depth (CD) of 0.75, illustrates a significant relationship between morphometric structure and live weight of cattle.