The relationship between Juhász-Nagy's information theory functions and the log-linear contingency table analysis

Abstract

Spatial pattern of the vegetation, due to its complexity, cannot be described by ad hoc indices. It requires a solid methodological basis, which can describe different levels of patterns (patterns of populations, coalitions and whole community) in the same framework. There are two methods, which seem to be appropriate for this purpose: Juhász-Nagy's information theory functions and log-linear contingency table analysis. This paper shows that from mathematical point of view they are close relatives. The main advantage of Juhász-Nagy's approach that it is developed to describe spatial pattern of the vegetation, therefore biological meaning of the functions is a central part of the approach. Whereas log-linear contingency table analysis is a general statistical method without any special (biological) meaning of terms. On the other hand, it is a well-known statistical method, while most of the biologists are unfamiliar in information theory. The relationship between these two approaches makes it possible to hybridise their advantages.