Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms

Anthony Tedde; Clement Grelet; Phuong N. Ho; Jennie E. Pryce; Dagnachew Hailemariam; Zhiquan Wang; Graham Plastow; Nicolas Gengler; Yves Brostaux; Eric Froidmont; Frédéric Dehareng; Carlo Bertozzi; Mark A. Crowe; Isabelle Dufrasne; GplusE Consortium; Hélène Soyeurt

doi:10.3390/ani11051288

Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms

Anthony Tedde^*, Clement Grelet, Phuong N. Ho, Jennie E. Pryce, Dagnachew Hailemariam, Zhiquan Wang, Graham Plastow, Nicolas Gengler, Yves Brostaux, Eric Froidmont, Frédéric Dehareng, Carlo Bertozzi, Mark A. Crowe, Isabelle Dufrasne, GplusE Consortium, Hélène Soyeurt

^*Corresponding author for this work

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaper › Journal article › Research › peer-review

13 Citations (Scopus)

Abstract

Knowing the body weight (BW) of a cow at a specific moment or measuring its changes through time is of interest for management purposes. The current work aimed to validate the feasibility of predicting BW using the day in milk, parity, milk yield, and milk mid-infrared (MIR) spectrum from a multiple-country dataset and reduce the number of predictors to limit the risk of over-fitting and potentially improve its accuracy. The BW modeling procedure involved feature selections and herd-independent validation in identifying the most interesting subsets of predictors and then external validation of the models. From 1849 records collected in 9 herds from 360 Holstein cows, the best performing models achieved a root mean square error (RMSE) for the herd-independent validation between 52 ± 2.34 kg to 56 ± 3.16 kg, including from 5 to 62 predictors. Among these models, three performed remarkably well in external validation using an independent dataset (N = 4067), resulting in RMSE ranging from 52 to 56 kg. The results suggest that multiple optimal BW predictive models coexist due to the high correlations between adjacent spectral points.

Original language	English
Article number	1288
Journal	Animals
Volume	11
Issue	5
Number of pages	15
ISSN	2076-2615
DOIs	https://doi.org/10.3390/ani11051288
Publication status	Published - May 2021

Keywords

dairy cow bodyweight
dairy cows
dimensionality reduction
feature selection
machine learning
mid infrared spectra
partial least square
Dimensionality reduction
Feature selection
Mid infrared spectra
Partial least square
Dairy cows
Machine learning
Dairy cow bodyweight

Access to Document

10.3390/ani11051288Licence: CC BY

Cite this

Tedde, A., Grelet, C., Ho, P. N., Pryce, J. E., Hailemariam, D., Wang, Z., Plastow, G., Gengler, N., Brostaux, Y., Froidmont, E., Dehareng, F., Bertozzi, C., Crowe, M. A., Dufrasne, I., GplusE Consortium, & Soyeurt, H. (2021). Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms. Animals, 11(5), Article 1288. https://doi.org/10.3390/ani11051288

@article{ba343c33e4674985af20a13499351324,

title = "Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms",

abstract = "Knowing the body weight (BW) of a cow at a specific moment or measuring its changes through time is of interest for management purposes. The current work aimed to validate the feasibility of predicting BW using the day in milk, parity, milk yield, and milk mid-infrared (MIR) spectrum from a multiple-country dataset and reduce the number of predictors to limit the risk of over-fitting and potentially improve its accuracy. The BW modeling procedure involved feature selections and herd-independent validation in identifying the most interesting subsets of predictors and then external validation of the models. From 1849 records collected in 9 herds from 360 Holstein cows, the best performing models achieved a root mean square error (RMSE) for the herd-independent validation between 52 ± 2.34 kg to 56 ± 3.16 kg, including from 5 to 62 predictors. Among these models, three performed remarkably well in external validation using an independent dataset (N = 4067), resulting in RMSE ranging from 52 to 56 kg. The results suggest that multiple optimal BW predictive models coexist due to the high correlations between adjacent spectral points.",

keywords = "dairy cow bodyweight, dairy cows, dimensionality reduction, feature selection, machine learning, mid infrared spectra, partial least square, Dimensionality reduction, Feature selection, Mid infrared spectra, Partial least square, Dairy cows, Machine learning, Dairy cow bodyweight",

author = "Anthony Tedde and Clement Grelet and Ho, {Phuong N.} and Pryce, {Jennie E.} and Dagnachew Hailemariam and Zhiquan Wang and Graham Plastow and Nicolas Gengler and Yves Brostaux and Eric Froidmont and Fr{\'e}d{\'e}ric Dehareng and Carlo Bertozzi and Crowe, {Mark A.} and Isabelle Dufrasne and {GplusE Consortium} and H{\'e}l{\`e}ne Soyeurt and Leslie Foldager and Torben Larsen and Tine Rousing and Jehan Ettema and S{\o}ren {\O}stergaard and S{\o}rensen, {Martin Tang} and Ingvartsen, {Klaus L{\o}nne}",

year = "2021",

month = may,

doi = "10.3390/ani11051288",

language = "English",

volume = "11",

journal = "Animals",

issn = "2076-2615",

publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",

number = "5",

}

Tedde, A, Grelet, C, Ho, PN, Pryce, JE, Hailemariam, D, Wang, Z, Plastow, G, Gengler, N, Brostaux, Y, Froidmont, E, Dehareng, F, Bertozzi, C, Crowe, MA, Dufrasne, I, GplusE Consortium & Soyeurt, H 2021, 'Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms', Animals, vol. 11, no. 5, 1288. https://doi.org/10.3390/ani11051288

Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms. / Tedde, Anthony; Grelet, Clement; Ho, Phuong N. et al.
In: Animals, Vol. 11, No. 5, 1288, 05.2021.

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaper › Journal article › Research › peer-review

TY - JOUR

T1 - Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms

AU - Tedde, Anthony

AU - Grelet, Clement

AU - Ho, Phuong N.

AU - Pryce, Jennie E.

AU - Hailemariam, Dagnachew

AU - Wang, Zhiquan

AU - Plastow, Graham

AU - Gengler, Nicolas

AU - Brostaux, Yves

AU - Froidmont, Eric

AU - Dehareng, Frédéric

AU - Bertozzi, Carlo

AU - Crowe, Mark A.

AU - Dufrasne, Isabelle

AU - GplusE Consortium

AU - Soyeurt, Hélène

AU - Foldager, Leslie

AU - Larsen, Torben

AU - Rousing, Tine

AU - Ettema, Jehan

AU - Østergaard, Søren

AU - Sørensen, Martin Tang

AU - Ingvartsen, Klaus Lønne

PY - 2021/5

Y1 - 2021/5

N2 - Knowing the body weight (BW) of a cow at a specific moment or measuring its changes through time is of interest for management purposes. The current work aimed to validate the feasibility of predicting BW using the day in milk, parity, milk yield, and milk mid-infrared (MIR) spectrum from a multiple-country dataset and reduce the number of predictors to limit the risk of over-fitting and potentially improve its accuracy. The BW modeling procedure involved feature selections and herd-independent validation in identifying the most interesting subsets of predictors and then external validation of the models. From 1849 records collected in 9 herds from 360 Holstein cows, the best performing models achieved a root mean square error (RMSE) for the herd-independent validation between 52 ± 2.34 kg to 56 ± 3.16 kg, including from 5 to 62 predictors. Among these models, three performed remarkably well in external validation using an independent dataset (N = 4067), resulting in RMSE ranging from 52 to 56 kg. The results suggest that multiple optimal BW predictive models coexist due to the high correlations between adjacent spectral points.

AB - Knowing the body weight (BW) of a cow at a specific moment or measuring its changes through time is of interest for management purposes. The current work aimed to validate the feasibility of predicting BW using the day in milk, parity, milk yield, and milk mid-infrared (MIR) spectrum from a multiple-country dataset and reduce the number of predictors to limit the risk of over-fitting and potentially improve its accuracy. The BW modeling procedure involved feature selections and herd-independent validation in identifying the most interesting subsets of predictors and then external validation of the models. From 1849 records collected in 9 herds from 360 Holstein cows, the best performing models achieved a root mean square error (RMSE) for the herd-independent validation between 52 ± 2.34 kg to 56 ± 3.16 kg, including from 5 to 62 predictors. Among these models, three performed remarkably well in external validation using an independent dataset (N = 4067), resulting in RMSE ranging from 52 to 56 kg. The results suggest that multiple optimal BW predictive models coexist due to the high correlations between adjacent spectral points.

KW - dairy cow bodyweight

KW - dairy cows

KW - dimensionality reduction

KW - feature selection

KW - machine learning

KW - mid infrared spectra

KW - partial least square

KW - Dimensionality reduction

KW - Feature selection

KW - Mid infrared spectra

KW - Partial least square

KW - Dairy cows

KW - Machine learning

KW - Dairy cow bodyweight

UR - http://www.scopus.com/inward/record.url?scp=85106548935&partnerID=8YFLogxK

U2 - 10.3390/ani11051288

DO - 10.3390/ani11051288

M3 - Journal article

SN - 2076-2615

VL - 11

JO - Animals

JF - Animals

IS - 5

M1 - 1288

ER -

Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

GplusE: GplusE - Genotype and Environment contributing to the sustainability of dairy cow production systems through the optimal integration of genomic selection and novel management protocols based on the development

Cite this

Validation of dairy cow bodyweight prediction using traits easily recorded by dairy herd improvement organizations and its potential improvement using feature selection algorithms

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Projects

GplusE: GplusE - Genotype and Environment contributing to the sustainability of dairy cow production systems through the optimal integration of genomic selection and novel management protocols based on the development

Cite this