Disentangling data dependency using cross-validation strategies to evaluate prediction quality of cattle grazing activities using machine learning algorithms and wearable sensor data.

Wearable sensors have been explored as an alternative for real-time monitoring of cattle feeding behavior in grazing systems. To evaluate the performance of predictive models such as machine learning (ML) techniques, data cross-validation (CV) approaches are often employed. However, due to data dependencies and confounding effects, poorly performed validation strategies may significantly inflate the prediction quality. In this context, our objective was to evaluate the effect of different CV strategies on the prediction of grazing activities in cattle using wearable sensor (accelerometer) data and ML algorithms. Six Nellore bulls (average live weight of 345 ± 21 kg) had their behavior visually classified as grazing or not-grazing for a period of 15 d. Elastic Net Generalized Linear Model (GLM), Random Forest (RF), and Artificial Neural Network (ANN) were employed to predict grazing activity (grazing or not-grazing) using 3-axis accelerometer data. For each analytical method, three CV strategies were evaluated: holdout, leave-one-animal-out (LOAO), and leave-one-day-out (LODO). Algorithms were trained using similar dataset sizes (holdout: n = 57,862; LOAO: n = 56,786; LODO: n = 56,672). Overall, GLM delivered the worst prediction accuracy (53%) compared with the ML techniques (65% for both RF and ANN), and ANN performed slightly better than RF for LOAO (73%) and LODO (64%) across CV strategies. The holdout yielded the highest nominal accuracy values for all three ML approaches (GLM: 59%, RF: 76%, and ANN: 74%), followed by LODO (GLM: 49%, RF: 61%, and ANN: 63%) and LOAO (GLM: 52%, RF: 57%, and ANN: 57%). With a larger dataset (i.e., more animals and grazing management scenarios), it is expected that accuracy could be increased. Most importantly, the greater prediction accuracy observed for holdout CV may simply indicate a lack of data independence and the presence of carry-over effects from animals and grazing management. Our results suggest that generalizing predictive models to unknown (not used for training) animals or grazing management may incur poor prediction quality. The results highlight the need for using management knowledge to define the validation strategy that is closer to the real-life situation, i.e., the intended application of the predictive model.

Replacement of raw soybean with roasted soybean increased milk production in Holstein cows / Substituição de soja crua por soja tostada aumentou a produção de leite em vacas da raça Holandesa

ABSTRACT: The aim of this study was to evaluate the effect of total replacement of raw whole soybean (RAW) for roastedwhole soybean (ROS) on the production performance of Holstein cows. Two experiments were carried out usinga simple reversal design where RAW has been completely replaced by ROS. In experiment 1, 22 cows (175±60 days in milk)were used, and the dietary inclusion level of RAW or ROS was 3.7% of dry matter (DM). In experiment 2, 16 cows (130±50 days in milk)were used, and thedietary inclusion level of RAW or ROS was 11% of DM. In both experiments, ROS increased milk production by 1.1kgday-1 without changing fat and protein production. Dry matter intake or milk urea nitrogenwere not affected by dietary soy source. In experiment 2, plasma glucose concentration was decreased, and allantoin/creatinine ratio in urine tended to decreasein ROS. Experiment 2 also evaluated the nutrient digestibility and ruminal degradation kinetics of crude protein in two soybean sources. Roasting had no effect on the digestibility of DM, organic matter, and neutral detergent fiber. Roasted whole soybean hadgreater fraction B and lower protein degradation rate than did RAW; this showed that heat treatment was effective in increasing therumen undegradable amino acid flowto the animal, which suggesteda potential mechanism of action for improved performance observed in ROS.

RESUMO: O objetivo deste estudo foi avaliar o efeito da substituição total de soja integral crua (SC) por soja integral tostada (ST) sobre o desempenho produtivo de vacas Holandês. Foram realizados dois experimentos com delineamento experimental de reversão simples nos quais a SC foi totalmente substituída por ST. No experimento 1, foram utilizadas 22 vacas (175±60DEL) e o nível de inclusão dietética de SC ou ST foi de 3,7% na matéria seca (MS). No experimento 2, foram utilizadas 16 vacas (130±50 DEL) e o nível de inclusão dietético de SC ou ST foi de 11% na MS. Em ambos os experimentos, ST aumentou a produção de leite em 1,1kgd-1, sem alterar as produções de gordura e proteína. O consumo de MS e nitrogênio uréico no leite não foram afetados pela fonte de soja dietética. No experimento 2, a concentração de glicose plasmática foi reduzida e a relação alantoína/creatinina na urina tendeu a ser reduzida por ST. O experimento 2 também avaliou digestibilidade de nutrientes e cinética de degradação ruminal da proteína bruta das duas fontes de soja. Não houve efeito da tostagem nas digestibilidades da MS, matéria orgânica e FDN. A ST apresentou maior fração B e menor taxa de degradação da proteína do que a SC, mostrando que o tratamento térmico foi efetivo em aumentar o fluxo de aminoácidosnão degradáveis no rúmen para o animal, sugerindo um potencial mecanismo de ação para a melhora no desempenho observada com ST.