Prediction of potato yield based on energy inputs using multi-layer adaptive neuro-fuzzy inference system

In this study two intelligent systems, based on adaptive neuro-fuzzy inference systems (ANFIS) and artificial neural networks (ANNs), were adapted to predict potato yield based on energy inputs. Data were collected from Isfahan province, Iran. Energy inputs included labor, machinery, diesel fuel, seeds, biocides, chemical fertilizers (N, P2O5 and K2O), farmyard manure, irrigation water and electricity. The best ANN model had a 11-30-2-1 structure, i.e., it consisted of an input layer with eleven input variables, two hidden layers with 30 and 2 neurons respectively, and potato yield as output. The best ANFIS model was designed using eight ANFIS sub-networks which were developed at three stages. Correlation coefficient (R), root mean square error (RMSE) and mean absolute percentage error (MAPE) for the best ANN model were computed as 0.925, 0.071 and 0.5, respectively. The corresponding R, RMSE and MAPE values for the best ANFIS topology were 0.987, 0.029 and 0.2, respectively. Based on the results of this study, it can be concluded that multi-layer ANFIS model due to employing fuzzy rules, gives better results than does ANN model.