Learning the Structure of Dynamic Probabilistic Networks

TLDR

Dynamic probabilistic networks compactly represent complex stochastic processes. The study aims to learn DPN structure from data and evaluate its use in predicting dynamic behaviors and uncovering causal orderings in biology. The authors extend scoring rules to dynamic networks and search for structure even when some variables are hidden. Empirical results demonstrate the methods’ applicability to both predictive/classification tasks and biological causal ordering.

Abstract

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.