Selection of linear mixed models to estimate the amount of zinc (KG/HA) on jalo bean

Abstract

The focus of the work was to develop a mixed linear model for estimating the amount of Zinc (kg/ha), considering two factors: time and planting Density. The selected model were included random effects intercept and time, beyond the ARH(1) covariance structure Autoregressive with heterogeneity of variances, managing and explain the correlations and heterogeneity of variances in different levels of planting Density. To compare the models used the criteria of AIC and BIC.