Conditional von Mises Bayesian Networks

Directed acyclic graphical models, or Bayesian networks, use a directed acyclic graph to represent the conditional independence relationships between a set of random variables. We introduce a novel class of Bayesian networks specifically designed for circular or angular variables, utilizing the properties of the von Mises distribution. We illustrate our proposal by applying these models to study the conditional independencies within a sequence of angles that characterize the structure of a glycopeptide.