Most of crop modeling is mechanistic model whereas the objective of this study was to predict mungbean growth and yield by artificial neural network (ANN). The experiment was 2 varieties (SUT1 and KPS2) x 2 water levels (rainfed and irrigation) x 3 fertilizer levels (12-24-12 rate 0, 15 and 30kg./rai) factorial experiment in RCBD 4 blocks growing for 2 seasons (no rainfed in the 2nd season). Data was collected for 10 replications in each plot. Using 560 sets of data from fertilizer rate of 0 and 30 kg/rai trained the ANN model by back propagation algorithm. The input variables were variety, fertilizer, irrigations, seasonal, growing degree day, rainfall, solar radiation and day after planting. The rest of 280 sets of data (only 15 kg/rai of fertilizer)... |