Classification of Daily Body Weight Gains in Beef Calves Using Decision Trees, Artificial Neural Networks, and...

Area of Science:

• Animal Science
• Machine Learning
• Predictive Analytics

Background:

• Accurate prediction of daily body weight gains in beef calves is crucial for optimizing livestock management and economic outcomes.
• Traditional methods may not fully capture the complex interplay of factors influencing calf growth.

Purpose of the Study:

• To compare the predictive performance of decision trees, artificial neural networks, and logistic regression for classifying daily body weight gains in beef calves.
• To identify key predictors influencing pre-weaning daily body weight gains in Simmental and Limousin breeds.

Main Methods:

• Utilized a dataset of 680 Simmental and 373 Limousin cows and their calves fattened between 2014 and 2016.
• Developed and compared machine learning models (decision trees, ANNs, logistic regression, random forest) for classifying daily weight gains into two categories (below or above breed/sex specific mean).
• Evaluated model performance using sensitivity, specificity, accuracy, and AUC on a test set.

Main Results:

• The random forest model demonstrated strong predictive performance for both Limousin (accuracy 0.76, AUC 0.82) and Simmental (accuracy 0.78, AUC 0.81) breeds.
• Significant predictors included dam's own growth history, previous calf performance, calf sex, dam's milk yield, birth weights, and dam's physical characteristics.

Conclusions:

• Machine learning models, particularly random forest, offer effective tools for classifying beef calf daily weight gains.
• The identified predictors provide valuable insights for breeding programs and early management interventions to enhance calf growth.