A0970
Title: Gaussian-related graphical models for circular variables
Authors: Anna Gottard - University of Firenze (Italy)
Agnese Panzera - University of Florence (Italy) [presenting]
Abstract: Graphical models are a key class of probabilistic models for studying conditional independence among random variables. Two classes of multivariate circular distributions related, albeit in different manners, to the Gaussian distribution are considered. The main properties of these distributions are explored in terms of conditional independence, proposing related classes of graphical models. The usefulness of the proposed models is illustrated by modelling the conditional independence among dihedral angles, which are crucial for defining the three-dimensional structure of proteins.
