Learning the Structure of Dynamic Probabilistic Networks

Dynamic probabilistic networks are a compact representation of complex stochastic processes. In this paper we examine how to learn the structure of a DPN from data. We extend structure scoring rules for standard probabilistic networks to the dynamic case, and show how to search for structure when some of the variables are hidden. Finally, we examine two applications where such a technology might be useful: predicting and classifying dynamic behaviors, and learning causal orderings in biological processes. We provide empirical results that demonstrate the applicability of our methods in both domains. 1 Introduction Probabilistic networks (PNs), also known as Bayesian networks or belief networks, are already well-established as representations of domains involving uncertain relations among several random variables. Somewhat less well-established, but perhaps of equal importance, are dynamic probabilistic networks (DPNs), which model the stochastic evolution of a set of random variables ..