Descrição da curva de crescimento de frutos do cajueiro por modelos não lineares

Abstract

The objective of this study was to evaluate the adjustment of gompertz and logistic models, with independent and autoregressive error structure in cashew fruits growth, based on the length and width measurements of the fruit, taken through the time. The estimation of the parameters was done in R software routines using the method of minimal squares and the interactive process of Gauss-Newton. The models were compared by the following criteria: the adjusted coefficient of determination, residual standard deviation, Akaike information criterion and Schwarz Bayesian criterion. For both models, the data showed positive residual autocorrelation for both variables: length and width of the fruit, described in autoregressive process of order one. In general, both models adjusted to the data, however there was an overestimation of the fruit size at the late stages. The logistic model was the most adequate to describe the cashew fruit growth in length and width.