Prediction of Wheat Production Using Artificial Neu ral Networks and Investigating Indirect Factors Affecti ng It: Case Study in Canterbury Province, New Zealand

Abstract

An artificial neural network (ANN) approach was use d to model the wheat production. From an extensive data collection involving 40 farm s in Canterbury, New Zealand, the average wheat production was estimated at 9.9 t ha -1 . The final ANN model developed was capable of predicting wheat production under differ ent conditions and farming systems using direct and indirect technical factors. After examining more than 140 different factors, 6 factors were selected as influential inp ut into the model. The final ANN model can predict wheat production based on farm conditio ns (wheat area and irrigation frequency), machinery condition (tractor hp ha -1 and number of passes of sprayer) and farm inputs (N and fungicides consumption) in Cante rbury with an error margin of ±9% (±0.89 t ha -1 ). Keywords : Agricultural Production, Agricultural Systems, Mo delling.