Comparison of linear regression and machine learning for predicting dry matter intake in Nellore cattle
Peer Review · 17.11.2026 · 8 (12) · 433-449
DOI: 10.53660/prw-3290-5620 ↗Kurz gefasst
Fachartikel in der Zeitschrift Peer Review, online erschienen am 17.11.2026. Verfasst von Alexandre Silva de Moraes und 3 weiteren Autor:innen.
Abstract (Originalsprache)
This study compared linear regression and machine learning algorithms (SVR, XGBoost, and MLP) for predicting dry matter intake (DMI) and evaluating residual feed intake (RFI) in Nellore cattle. Data from 1,381 bulls across nine performance tests (2016–2024) were analyzed using average daily gain and metabolic body weight as predictors. Models were evaluated through within-test cross-validation and leave-one-test-out (LOTO) validation. Under LOTO, MLP achieved the lowest RMSE (1.3590 kg/day), followed by linear regression (1.3725 kg/day), SVR (1.3745 kg/day), and XGBoost (1.3833 kg/day). Despite systematic prediction biases across tests, agreement with the traditional Koch method in selecting the 20% of animals with the lowest RFI within each test ranged from 75.91% to 82.48%. Machine learning provided no substantial predictive advantage over linear regression, which represents a simpler and competitive approach for DMI prediction under the conditions evaluated.
Zitieren
APA
Moraes, A. S. D., Menezes, G. R. D. O., Toral, F. L. B., & Araújo, C. V. D. (2026). Comparison of linear regression and machine learning for predicting dry matter intake in Nellore cattle. Peer Review, 8(12), 433-449. https://doi.org/10.53660/prw-3290-5620
Vancouver
Moraes ASD, Menezes GRDO, Toral FLB, Araújo CVD. Comparison of linear regression and machine learning for predicting dry matter intake in Nellore cattle. Peer Review. 2026;8(12):433-449. doi:10.53660/prw-3290-5620
BibTeX
@article{Moraes2026,
author = {Moraes, Alexandre Silva de and Menezes, Gilberto Romeiro de Oliveira and Toral, Fabio Luiz Buranelo and Araújo, Cláudio Vieira de},
title = {Comparison of linear regression and machine learning for predicting dry matter intake in Nellore cattle},
journal = {Peer Review},
year = {2026},
volume = {8},
number = {12},
pages = {433--449},
doi = {10.53660/prw-3290-5620}
}
EUVIMED