Métodos não paramétricos em processos pontuais replicados marcados

Abstract

In several areas of knowledge, it is of great significance, to analyze the spatial dependency structure of a random variable. Therefore, the aim of this thesis is to present statistical methods to analyze the structure of spatial dependence in data of marked and unmarked point patterns. By this way, the work presents several methods of point patterns in the plan; those are consolidated in the literature and propose a new method to test the equality hypothesis between groups of marked (quantitative) replicated point process. The suggested method is evaluated by performing type I error analysis. All presented methods are applied to characterize the spatial distribution of alveoli located in the frozen dough under different temperatures of French type bread. The obtained results allow us to affirm that the methods of replicated point patterns indispensable tools characterize the spatial dependence of events located in a region.