Comparison of MLP and RBF neural networks performance for estimation of broiler output energy

Energy management is one of the main ways of the efficient use of energy resources. The prediction of crop yields based on energy inputs can help farmers and policymakers to estimate the level of production. Required data for study were randomly collected from 70 broiler farms in North West of Iran. The input energies were included human labour, machinery, fuel, feed and electricity and the output produced energies were considered as output variables. The multi-layer perceptron (MLP) and the radial basis function (RBF) neural networks were applied for prediction of output energies of broiler production. According to the comparison results obtained from the indices of the coefficient of determination (R2), root mean square error (RMSE) and the mean absolute error (MAE) performance of the ANN-RBF model better than ANN-MLP model was estimated. In evaluating the effects of inputs on outputs of production, the production of fossil fuel showed the highest sensitivity among the production inputs in both models.