Measures of Dependence Structure for Random Sets (1998)

Abstract:

A set of measures of dependence are proposed for the analysis of the structure of random sets. These measures are similar to cumulants in some aspects, but are far more suited to the structure of random sets than cumulants. Differences were found between cumulants and -the new functions in terms of probabilistic, statistical and geometrical criteria. In each case, the new measures were found to be superior to ordinary cumulants in the context of this particular type of stochastic process. 

Keywords:

cumulants, random sets, germ-grain model, Boolean Model, Dead Leaves Model AMS Subject Classifications: 60D05