Prediction of yield and economic indices for tangerine production using artificial neural networks based on energy consumption

Abstract Determination of suitable model for forecasting of yield and economic indices of tangerine production in Guilan province of Iran using artificial neural network (ANN) was the main aim of this study. For this purpose, the energy consumption for three groups size of tangerine orchards were found from 60 questionnaires. The results revealed the average total energy use and yield of tangerine production were 27873 MJ ha-1 and 25740 kg ha-1, respectively. In the next step, the economic indices were calculated for tangerine orchards. Accordingly, benefit to cost ratio, productivity, net return and energy intensiveness were calculated as 1.37, 3.42 kg $-1, 2777.82 $ ha-1, 2.71 $ ha-1, respectively. In this study, a back propagation algorithm was used for training of ANN model and Levenberg-Marquardt was learning algorithm. The best topology had the 10-8-5 structure. Moreover, the R2 of best structure was found 0.971, 0.954, 0.983, 0.991 and 0.973 for tangerine yield, benefit to cost ratio, productivity, net return and energy intensiveness, respectively. In the last section of this research, sensitivity analysis was done and results illustrated the highest sensitivity rate of tangerine yield, the benefit to cost ratio, productivity, net return and energy intensiveness was belonged to farmyard manure, insecticide, insecticide, phosphate and diesel fuel, respectively.