In this study, evaluate the control models to describe the profile of the Corymbia
citriodora. The data used are obtained from a commercial plant located in the
southern region of the state of Tocantins. There are data from 24 Corymbia
citriodora trees accurately cubed with diameter measurements in stem families:
0.2; 0.4; 0.7; 1,3; 2.70 m and, successively, from 2 m in 2 m to a diameter of
around 3 cm in shell. 44 taper models were tested, being 25 linear and 19 nonlinear.
In the evaluation of these models, we consider an analysis of the
distribution of residues, and statistics: standard error of estimation, coefficient of
determination, randomized block design with split plot arrangement, mean
absolute deviation, square root of mean error, sum of squares of residuals,
standard deviation of the differences, Akaike's information criterion. The bestperforming
model for an exponent-shape fit type adjusted in nonlinear form.
This version has a long list of corrections, changes, and new developments that
significantly improve functioning. In this comparison, statistical criteria are used:
mean deviation, bias, linear correlation, standard error of the estimate, between
observed and estimated values, precision, obtained by square test and square
root of error, as well as analysis of distribution and residues. It was concluded
that the use of a cube simulated by the method of elevation of the estimate, as
well as a conciliation equation as one obtains by the use of real data.
|
