Machine learning approaches for predicting feed intake in Australian Merino, Corriedale, and Dohne Merino sheep. (Original research article).

ABSTRACT.- Feed intake is a challenging trait to measure due to the high costs associated with labor, feeding, and facilities. Applying machine learning approaches, considering traits as potential predictors, offers a cost-effective alternative to direct feed intake measurement. By leveraging existi...

Descripción completa

Detalles Bibliográficos
Autores: AMARILHO-SILVEIRA, F., DE BARBIERI, I., NAVAJAS, E., COBUCI, J. A., CIAPPESONI, G.
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2025
País:Uruguay
Institución:Instituto Nacional de Investigación Agropecuaria
Repositorio:AINFO
Idioma:inglés
OAI Identifier:oai:redi.anii.org.uy:20.500.12381/5123
Acceso en línea:https://ainfo.inia.uy/consulta/busca?b=pc&id=65229&biblioteca=vazio&busca=65229&qFacets=65229
Access Level:acceso abierto
Palabra clave:K-nearest neighbor
Enteric methane
Carbon dioxide
Random forest
Support vector machines
SISTEMA GANADERO EXTENSIVO - INIA
Descripción
Sumario:ABSTRACT.- Feed intake is a challenging trait to measure due to the high costs associated with labor, feeding, and facilities. Applying machine learning approaches, considering traits as potential predictors, offers a cost-effective alternative to direct feed intake measurement. By leveraging existing animal data, these models can optimize resources and enable feed intake estimation across a larger population without the need for labor-intensive trials. This research aimed to test combinations offeature selection and prediction models to find the best feed intake (expressed as metabolizable energy intake) prediction approach for a dataset comprising AustralianMerino, Corriedale, and Dohne Merino data. The study dataset with 1,708 observations included 920 Australian Merino, 215 Corriedale, and 337 Dohne Merino sheep from 17 feed intake trials conducted between 2019 and 2022. The dataset was randomly partitioned into two subsets: one for training (80%) the algorithms and the other for direct validation (20%). © 2025 Amarilho-Silveira, De Barbieri, Navajas, Cobuci and Ciappesoni.