Australian Society of Information and Communication Technologies in Agriculture
Faculty of Health, Engineering and Science
School of Computer and Security Science/eAgriculture Research Group
This paper examines the application of artificial neural networks (ANNs) for predicting crop yields for an agricultural region in Nepal. The neural network algorithm has become an effective data mining tool and the outcome produced by this algorithm is considered to be less error prone than other computer science techniques. The backpropagation algorithm which iteratively finds a suitable weight value is considered for computing the error derivative. Agricultural data was collected from thirteen years from paddy field cultivation in the Siraha district, an eastern region in Nepal, and used for this investigation of neural networks. Additionally, climatic parameters including rainfall, maximum temperature and minimum temperature along with the fertilizer use were also used as input values. The experiment shows that the trained neural network produced a minimum error which indicated that the test model is capable of predicting crops yield in Nepal.