Prediction of Body Weight by Using PCA-Supported Gradient Boosting and Random Forest Algorithms in Water Buffaloes (Bubalus bubalis) Reared in South-Eastern Mexico.

This study aims to use advanced machine learning techniques supported by Principal Component Analysis (PCA) to estimate body weight (BW) in buffalos raised in southeastern Mexico and compare their performance. The first stage of the current study consists of body measurements and the process of determining the most informative variables using PCA, a dimension reduction method. This process reduces the data size by eliminating the complex structure of the model and provides a faster and more effective learning process. As a second stage, two separate prediction models were developed with Gradient Boosting and Random Forest algorithms, using the principal components obtained from the data set reduced by PCA. The performances of both models were compared using R2, RMSE and MAE metrics, and showed that the Gradient Boosting model achieved a better prediction performance with a higher R2 value and lower error rates than the Random Forest model. In conclusion, PCA-supported modeling applications can provide more reliable results, and the Gradient Boosting algorithm is superior to Random Forest in this context. The current study demonstrates the potential use of machine learning approaches in estimating body weight in water buffalos, and will support sustainable animal husbandry by contributing to decision making processes in the field of animal science.

Predicting carcass tissue composition in Blackbelly sheep using ultrasound measurements and machine learning methods.

This study aimed to predict Blackbelly sheep carcass tissue composition using ultrasound measurements and machine learning models. The models evaluated were decision trees, random forests, support vector machines, and multi-layer perceptrons and were used to predict the total carcass bone (TCB), total carcass fat (TCF), and total carcass muscle (TCM). The best model for predicting the three parameters, TCB, TCF, and TCM was random forests, with mean squared error (MSE) of 0.31, 0.33, and 0.53; mean absolute error (MAE) of 0.26, 0.29, and 0.53; and the coefficient of determination (R2) of 0.67, 0.69, and 0.76, respectively. The results showed that machine learning methods from in vivo ultrasound measurements can be used as determinants of carcass tissue composition, resulting in reliable results.

Predicting live weight using body volume formula in lactating water buffalo.

Live weight (LW) is an important piece of information within production systems, as it is related to several other economic characteristics. However, in the main buffalo-producing regions in the world, it is not common to periodically weigh the animals. We develop and evaluate linear, quadratic, and allometric mathematical models to predict LW using the body volume (BV) formula in lactating water buffalo (Bubalus bubalis) reared in southeastern Mexico. The LW (391.5 ± 138.9 kg) and BV (333.62 ± 58.51 dm3) were measured in 165 lactating Murrah buffalo aged between 3 and 10 years. The goodness-of-fit of the models was evaluated using the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), coefficient of determination (R2), mean-squared error (MSE) and root MSE (RMSE). In addition, the developed models were evaluated through cross-validation (k-folds). The ability of the fitted models to predict the observed values was evaluated based on the RMSEP, R2, and mean absolute error (MAE). LW and BV were significantly positively and strongly correlated (r = 0.81; P < 0.001). The quadratic model had the lowest values of MSE (2788.12) and RMSE (52.80). On the other hand, the allometric model showed the lowest values of BIC (1319.24) and AIC (1313.07). The Quadratic and allometric models had lower values of MSEP and MAE. We recommend the quadratic and allometric models to predict the LW of lactating Murrah buffalo using BV as a predictor.

Models to predict live weight from heart girth in crossbred beef heifers.

The objective of this study was to develop and evaluate linear, quadratic, and exponential mathematical models to predict live weight (LW) from heart girth (HG) in crossbred heifers raised in tropical humid conditions in Mexico. Live weight (363.32 ± 150.88 kg) and HG (166.83 ± 24.88 cm) were measured in 400 heifers aged between 3 and 24 months. Linear and non-linear regression was used to construct the prediction models. The goodness of fit of the models was evaluated using the Akaike information criterion (AIC), the Bayesian information criterion (BIC), coefficient of determination (R2), mean squared error (MSE), and root MSE (RMSE). In addition, the developed models were evaluated through internal and external cross-validation (k-folds) using independent data. The ability of the fitted models to predict the observed values was evaluated based on the root mean square error of prediction (RMSEP), R2, and mean absolute error (MAE). The correlation coefficient between LW and HG was r = 0.98 (P < 0.001). The quadratic model showed the lowest values of MAE (736.57), RMSEP (27.13), AIC (3783.95), and BIC (3799.91). Additionally, this model exhibited better goodness-of-fit values regarding external and internal validation criteria (higher R2 and lower RMSEP and MAE), thus having better predictive performance. The RMSE represented about 8% of the observed LW. Heart girth is highly correlated (r = 0.98) with LW. The quadratic model showed a high predictive capacity for crossbred heifers kept in tropical conditions.

Towards the Estimation of Body Weight in Sheep Using Metaheuristic Algorithms from Biometric Parameters in Microsystems.

The Body Weight (BW) of sheep is an important indicator for producers. Genetic management, nutrition, and health activities can benefit from weight monitoring. This article presents a polynomial model with an adjustable degree for estimating the weight of sheep from the biometric parameters of the animal. Computer vision tools were used to measure these parameters, obtaining a margin of error of less than 5%. A polynomial model is proposed after the parameters were obtained, where a coefficient and an unknown exponent go with each biometric variable. Two metaheuristic algorithms determine the values of these constants. The first is the most extended algorithm, the Genetic Algorithm (GA). Subsequently, the Cuckoo Search Algorithm (CSA) has a similar performance to the GA, which indicates that the value obtained by the GA is not a local optimum due to the poor parameter selection in the GA. The results show a Root-Mean-Squared Error (RMSE) of 7.68% for the GA and an RMSE of 7.55% for the CSA, proving the feasibility of the mathematical model for estimating the weight from biometric parameters. The proposed mathematical model, as well as the estimation of the biometric parameters can be easily adapted to an embedded microsystem.

Using Post-Mortem Measurements to Predict Carcass Tissue Composition in Growing Rabbits.

The objective of this study was to determine post-mortem measurements for predicting carcass traits in growing rabbits. A total of 50 clinically healthy New Zealand White × Californian male rabbits with a body weight (BW) of 1351 ± 347 g between 60 to 80 days of age were used. Body weight was recorded 12 h before slaughtering. Data recorded at slaughtering included carcass weights (HCW). After cooling at 4 °C for 24 h, carcasses were weighed (CCW) and then were carefully split longitudinally with a band saw to obtain left and right halves. In the right half carcass, the following measurements were recorded using a tape measure: dorsal length (DL), thoracic depth (TD), thigh length (TL), carcass length (CL), lumbar circumference (LC). The compactness index (CCI) was calculated as the CCW divided by the CL. Thereafter, the right half carcass was weighed and manually deboned to record weights of muscle (TCM), and bone (TCB). The CCI explained of 93% of variation for TCM (R2 = 0.93 and a CV = 9.30%). In addition, the DL was the best predictor (p < 0.001) for TCB (R2 = 0.60 and a CV = 18.9%). Our results indicated that the use of carcass measurements could accurately and precisely (R2 = ≥ 0.60 and ≤0.95) be used as alternatives to predict the carcass tissues composition in growing rabbits.

Characterization of oxygen transfer in a 24-well microbioreactor system and potential respirometric applications.

The assessment of microbial processes is often done in Microbioreactor systems (MBRs), which allow for parallel cultivation in multiple independent wells. MBRs often include dissolved oxygen sensors, which are convenient for process characterization through oxygen uptake rate and other respirometric determinations. In order to assess respirometric potential of MBRs, a complete assessment of the DO fluorescent quenching sensors was done, showing that they presented a typical error of 0.56%, a signal to noise ratio of 189, a response time from 5.7 to 7.2 s and no drift over a period of 24 h. Then, KLa in the MBR was measured with different cassette and cap designs, liquid volumes, agitation rates, gas flow rates, temperatures and ionic strengths. KLa ranged from 8 to 90 h(-1), with a standard deviation between replicates from 2.8 to 17.5%. From these results and a numerical simulation, it was shown that the MBR tested allow the determination of oxygen uptake rates in a range from 0.038 to 3390 mg L(-1) h(-1), with a determination error less than 15%. Besides OUR determination, it was concluded that the MBR tested is also a convenient tool for dynamic pulse respirometry methods, based on experimental confirmation with four different cultures.