Estimación del rendimiento por hectárea de maíz duro seco en el Ecuador mediante redes neuronales artificiales

Abstract

Maize is an important grass in the economy and food security due to its use in human food and as a feed in the livestock industry in Ecuador, which motivates the study of the factors that influence the yield of this crop. The objective of this work is to determine a model of artificial neural networks to estimate the yield of maize per hectare for each production cycle, winter and summer, through the use of climatic variables, fertilization, geographic location of the crop and variety of the seed. For this, the influential factors in crop yield are determined for each cycle and a multiple linear regression model and a neural network model are proposed. Finally, the results obtained in both models are compared using goodness-of-fit statistics.