ABSTRACT
Crop productivity prediction techniques assist with adjusting for potential agronomic problems during the growing season. Several authors have reported that there is a correlation between leaf chlorophyll (Chl) content and yield. This study developed independent artificial neural network (ANN) models for soybean and corn in order to predict the crops' productive potentials using their respective yields and leaf Chl content data, measured at three stages of plant development. The ANN was deemed ready for testing through verification of the mean squared error and the number of epochs while training the neural network. While the model obtained when Chl was measured in the V6 stage of development explained more than 50% of the productivity data in corn, the models obtained for soybean did not explain more than 10% of the observed data. Attempts to improve the model through changes of the architecture of the neural network did not show any improvement in model.
Acknowledgments
The authors are grateful to the State University of Western Paraná, the Federal University of Technology–Paraná, the Araucária Foundation (Fundação Araucária), the Coordination for the Improvement of Higher Education Personnel (CAPES), and the National Council for Scientific and Technological Development (CNPq) for the support received. Also, the authors thank Erin Kiser for her contribution during English review.