Prediction for Individual Growth in a Random Environment

Abstract

In the literature it is usual to find classic regression models to describe the dynamics of a certain growth phenomenon. However, in phenomena of a dynamic nature, it is more appropriate to use models that are able to incorporate the dynamics of the growth process and the effect produced by the environmental random fluctuations on such dynamics. This can be done using stochastic differential equations (SDE) models. In this chapter, we start by comparing the quality of fitting and prediction using nonlinear regression models and SDE models. For the SDE models, we discuss the computation of asymptotic confidence intervals for prediction using simulation and the delta method. We show an application using cattle weight data from several females of the Mertolengo cattle breed.