ESTIMATION OF EUCALYPTUS HEIGHT BY MEANS OF NONLINEAR REGRESSION AND ARTIFICIAL NEURAL NETWORKS

Main Article Content

Diogo Guido Streck VENDRUSCOLO
Ronaldo DRESCHER
Hudson Santos SOUZA
Jhonny Pinto Vieira Mendes MOURA
Fernanda Meyer Dotto MAMORÉ
Tiago Altobelle da Silva SIQUEIRA

Abstract

In this work was evaluated the nonlinear regression modeling and artificial neural networks to estimate the eucalyptus trees tall. The stands are located in Itiquira, MT. Data were obtained from 23 parcels measured at four years old. The database was divided into two sets, one for the adjustment of regression models and training the networks (70%) and the other to validate the regression and validating the trained networks (30%). The nonlinear regression model was used the Gompertz and the networks of the multilayer perceptron (MLP). The evaluation of the estimates accuracy was by the multiple correlation coefficient between the observed and estimated heights, square root of the average error in percentage and graphical analysis. Both evaluated techniques are effective for predicting the height of trees. However, neural networks statistical criteria showed slightly higher relative to the regression.

Article Details

How to Cite
VENDRUSCOLO, D. G. S., DRESCHER, R., SOUZA, H. S., MOURA, J. P. V. M., MAMORÉ, F. M. D., & SIQUEIRA, T. A. da S. (2015). ESTIMATION OF EUCALYPTUS HEIGHT BY MEANS OF NONLINEAR REGRESSION AND ARTIFICIAL NEURAL NETWORKS. Brazilian Journal of Biometrics, 33(4), 556–569. Retrieved from https://biometria.ufla.br/index.php/BBJ/article/view/33
Section
Articles

Most read articles by the same author(s)