| Popis: |
The objectives of this study were to evaluate the net energy partition patterns of growing-finishing pigs at different growing stages and to develop the corresponding prediction models using non-linear regression and artificial neural networks. Twenty-four pigs with an initial body weight of ~ 30 kg were kept in metabolic cages and fed ad libitum and were moved into 6 respiration chambers in turns until ~ 90 kg. The net energy partition patterns, i.e. net energy for maintenance, net energy retained as protein, and net energy retained as lipid, were calculated based on indirect calorimetry and nitrogen balance techniques. The energy balance data collected through the animal trial was then randomly split into a training dataset containing 75% of the samples and a testing dataset containing the remaining 25% of the samples. The non-linear regression models and a series of artificial neural networks models were established on the training dataset to predict the metabolizable energy intake, net energy intake, net energy for maintenance, net energy retained as protein, and net energy retained as lipid of pigs. The best-fitted artificial neural networks models were selected by 5-fold cross-validation in the training dataset. The prediction performance of the best-fitted non-linear regression and artificial neural networks models were compared on the testing dataset. The results showed that the average net energy intakes of pigs were 17.71, 23.25, 24.56, and 28.96 MJ/d in 30-45 kg, 45-60 kg, 60-75 kg, and 75-90 kg, respectively. The net energy for maintenance and net energy retained as lipid (MJ/d) kept increasing as body weight increased from 30 kg to 90 kg, while the net energy retained as protein increased to its maximum value and then kept in a certain range of 4.64-4.88 MJ/d. The proportion of net energy for maintenance for pigs at 30-90 kg stayed within the range of 42.0-48.6%, while the proportion of net energy retained as lipid kept increasing. For the prediction models built based on the animal trial, artificial neural networks models exhibited better performance than non-linear regression models for all the target outputs. In conclusion, net energy partition patterns changed in different growth stages of pigs, and artificial neural networks models are more flexible and powerful than non-linear regression models in predicting the net energy partition patterns of growing-finishing pigs. |
