Adjustment of growth curves for the production of sugarcane

Abstract

The modeling of growth curves is very common in data from repeated measures in time, because it is possible to establish a functional relationship between the time and the treatments used. The main objective of this work is to obtain a functional relationship that establishes a polynomial association between the production, in tons, and the 5 years in which 3 sugarcane genotypes had their production measured in an experiment installed in the Puerto Rico Plant, in Alagoas, conducted by UFSC. The experimental design was the randomized blocks, with 4 replicates. In these models we assume fixed mean curves, around which there are random variations for the individual curves of the respective experimental units. It was verified by the AIC and BIC criteria that the best structure is the variance component (VC) to represent the covariance matrix of the experimental and unstructured error (UN) for the covariance matrix of the random effects parameters and that the polynomial of 3rd degree was selected to represent the average relation between production and the 5 years evaluated for the three genotypes.