PSVI-41 Application of Artificial Neural Network to Predict Physiological Stress Responses in Goats due to Transportation.

摘要

This study was conducted to determine if artificial neural network (ANN) can be used as a nonlinear modeling technique to more accurately predict physiological stress responses in goats due to transportation compared with statistical regression. Data on stress responses were obtained from experiments conducted at Fort Valley State University. Prediction models were developed for plasma cortisol and glucose concentrations, creatine kinase (CK) activity, neutrophil (N) and lymphocyte (L) counts, and N:L ratio as a function of time (0, 1, 2, 3, and 4 h; n = 16 goats/time) after transportation (input 1) and stocking density (25 vs. 50 goats) during transportation (input 2). Each response (output) data set was subdivided into training, testing, and validation sets. The NeuralWorks Predict Software and SAS were used to develop ANN and regression models, respectively. ANN was applied using the backpropagation (BP) and Kalman filter (KF) learning rules to develop nonparametric models, and Pearson correlation (R), standard deviation, bias factor, root mean square residual, and accuracy factor were used to measure the prediction performance of each model. Prediction plots for each physiological response indicated that ANN with BP learning rule had better accuracies compared with KF or regression models. Correlations between predicted and observed values were better with ANN-BP models with R-values of 0.87, 0.67, 0.50, 0.27, 0.42, and 0.50 for cortisol, glucose, CK, N, L, and N:L, respectively, than with regression models (R-values = 0.82, 0.50, 0.45, 0.13, 0.29, and 0.12). Sensitivity analysis revealed that stocking density did not affect predicted values. The results showed that the ANN models can predict responses more accurately compared with statistical regression despite high biological variabilities associated with animal stress. Because of its inherent ability to interpolate unseen patterns, ANN could be an effective tool in predicting stress responses in farm animals.