Modelo de von Bertalanffy em diferentes aplicações

Abstract

In nature, there are several phenomena that researchers are interested in studying and understanding how their development was, absorbing the greatest possible degree of information. In this context, statistical modeling is often used, and linear or non-linear regression are the most used. Although non-linear regression is more complex due to parameter estimation issues, it has the advantage of more parsimonious models because, in general, they have a lower number of parameters with practical or biological significance. In this study, different applications of the non-linear von Bertalanffy model were explored, which can use different allometric coefficients or parameterizations. The von Bertalanffy model was initially adopted to describe the growth of fish and crustaceans, however, this model is currently being used in several areas, from agricultural sciences to biology. The first application proposed here was in the description of the growth of crossbred rabbits using different data collection methodologies, longitudinal and cross-sectional, in which the first is a collection of the weight of the same animal over time and the second a collection of the variable under study (weight) in a single day, using animals of the same breed of different ages. The second application was the use of the non-linear von Bertalanffy model in the description of ammonia volatilization in conventional and enhanced efficiency nitrogen fertilizers applied to coffee plants. For the third application, non-linear quantile regression techniques and analysis of possible outliers were employed as a solution for deviations from the assumptions of the residuals of the fit of the von Bertalanffy model for the description of ammonia volatilization in conventional and enhanced efficiency nitrogen fertilizers applied to the coffee tree.