Answers are grounded in this paper and cite their sources.
Bovine respiratory disease is widespread, costly to the industry and has lifelong impacts. Its detection can be time consuming, and it is difficult to detect early. This study evaluated machine learning techniques using integrated data to accurately predict bovine respiratory disease in pre‑weaned calves. A total of 118 calves were monitored during their pre-weaning period (9 days ± 0.2 to 54 days ±0.3) with 100 calves included in the dataset. A newly developed management ear-tag for calf, alongside automatic feeder data, pedometer and daily environmental information were used against a gold standard daily health assessment (Wisconsin score), to build a bovine respiratory disease detection model. This study explored different machine learning modelling techniques (random forest, K-nearest neighbours, support vector machine, catboost) as well as paired-combinations of their outputs. To assess the efficacy of the modelling techniques two metrics were used: accuracy of predictions (alerts generated from the model) and accuracy of disease detection (correctly identified disease events). Individually, support vector machine and catboost were found to provide the best balance between these two metrics where 56 % to 60 % of the alerts aligned with a disease day and 56 % to 62 % of the disease events were correctly identified. The best-balanced combination was catboost and support vector machine (56 % of correct alerts and 76 % correctly identified disease events), but highest results were obtained by different techniques and combinations. This study demonstrated that the integration of multiple data sources and different machine learning techniques provide a viable early disease detection model in calves.
No takes yet. Share an insight, caveat, or question.
Brans et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: