AI-Based Predictive Modelling and Alert Framework for Mortality Risk and Cost-Benefit Analysis in Rabbit Production.

Area of Science:

• Animal Science
• Machine Learning
• Agricultural Economics

Background:

• Commercial rabbit production faces significant economic losses due to mortality events.
• Early identification of elevated mortality risk at the group level is crucial for mitigation.

Purpose of the Study:

• To develop and evaluate a machine learning-based framework for predicting rabbit mortality risk.
• To integrate predictive modeling with an alert system for decision support in commercial rabbitries.

Main Methods:

• Utilized routinely collected group-level production data.
• Developed and validated models using an XGBoost algorithm and StratifiedGroupKFold cross-validation.
• Translated predictions into an alert system prioritizing sensitivity.

Main Results:

• The XGBoost model demonstrated a recall of 0.78, precision of 0.59, and ROC-AUC of 0.72.
• The alert system maintained a moderate false alert rate.
• Cost-benefit analysis showed positive economic returns under moderate to optimistic intervention effectiveness assumptions.

Conclusions:

• The developed framework is feasible for predicting mortality risk and supporting decision-making in rabbit production.
• Real-world validation is necessary to confirm practical effectiveness under commercial farm conditions.
• Predictive modeling offers a valuable tool for enhancing biosecurity and economic viability in livestock farming.