Skip to main content
ARS Home » Research » Publications at this Location » Publication #76058
Title: NEURAL NETWORK MODEL FOR PREDICTING CORN YIELD

Author
item BODUR, SABAHATTIN - NORTH DAKOTA STATE UNIV.
item PANIGRAHI, SURANJAN - NORTH DAKOTA STATE UNIV.
item Colvin, Thomas

Submitted to: American Society of Agricultural Engineers Meetings Papers
Publication Type: Abstract Only
Publication Acceptance Date: 9/28/1996
Publication Date: N/A
Citation: N/A

Interpretive Summary:

Technical Abstract: A yield prediction model for corn using neural network was developed. This paper discusses the capability of neural networks for predicting corn yield and compare its performance with the fuzzy logic model developed by Ambuel. Three different neural networks, Back propagation, Radial Basis Function Network, and Modular Neural Network were evaluated. Organic matter, drought and ponding potential, monthly rainfall and temperature were the model inputs, and yield was the output of the model. Out of these three networks, Back propagation provided the best results with 69% test accuracy and 84% training accuracy. With additional selected input parameters, Radial Basis Function Network worked the best with 64% test and 87% training accuracy.