ARTIFICIAL NEURAL NETWORKS FOR PREDICTING SUGARCANE STALK AND TOTAL BIOMASS YIELD BASED ON MICRONUTRIENT RATES APPLIED IN THE PLANTING FURROW AND TO THE LEAVES

ARTIFICIAL NEURAL NETWORKS FOR PREDICTING SUGARCANE STALK AND TOTAL BIOMASS YIELD BASED ON MICRONUTRIENT RATES APPLIED IN THE PLANTING FURROW AND TO THE LEAVES

ABSTRACT Sugarcane is pivotal in the global bioeconomy, providing a renewable resource for products such as ethanol, sugar, bioenergy, animal feed, and bioplastics. Its versatility makes it an essential crop for industries seeking sustainable alternatives to fossil fuels. This study presents an adva...

Full description

Saved in:
Bibliographic Details
Main Authors: Alfredo Bonini Neto, Maikon V. da Silva Lira, Guilherme C. Meirelles, Luiz F. de M. Santos, Carolina dos S. B. Bonini, Reges Heinrichs
Format: Article
Language:English
Published: Sociedade Brasileira de Engenharia Agrícola 2025-04-01
Series:Engenharia Agrícola
Subjects:
artificial intelligence
predicting
fertilization
micronutrients
sugarcane
Online Access:http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0100-69162025001000301&lng=en&tlng=en
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:ABSTRACT Sugarcane is pivotal in the global bioeconomy, providing a renewable resource for products such as ethanol, sugar, bioenergy, animal feed, and bioplastics. Its versatility makes it an essential crop for industries seeking sustainable alternatives to fossil fuels. This study presents an advanced approach that uses artificial neural networks (ANNs), specifically a multilayer perceptron model, to accurately estimate sugarcane productivity and biomass. The model incorporates the effects of micronutrient applications, both in the planting furrow and on the leaves, effectively capturing the complex interactions that influence crop yield. During training, the ANN demonstrated high precision, achieving a mean squared error (MSE) of 0.000097 and an R2 of 0.98, closely aligning the predicted outputs with experimental results. In the validation phase, using previously unseen data, it maintained strong performance, with an MSE of 0.0008796. This performance supports the model's ability to generalize beyond the training set, reliably estimating sugarcane yield and biomass under varying conditions. These findings highlight the potential of ANN-based approaches to enhance agricultural management, offering a robust tool to optimize crop performance and improve resource allocation in real-world farming scenarios.
ISSN:0100-6916

Similar Items