A computational scheme for Reasoning in Dynamic Probabilistic Networks

TL;DR

This paper introduces a computational scheme for reasoning in dynamic probabilistic networks, enabling efficient inference, smoothing, and forecasting for complex, non-linear, multivariate systems with intricate independence structures.

Contribution

It generalizes classical time-series inference methods to handle complex dynamic systems within a probabilistic network framework.

Findings

Implemented on the HUGIN shell for practical use.

Provides efficient backward smoothing techniques.

Enables approximate forecasting for dynamic systems.

Abstract

A computational scheme for reasoning about dynamic systems using (causal) probabilistic networks is presented. The scheme is based on the framework of Lauritzen and Spiegelhalter (1988), and may be viewed as a generalization of the inference methods of classical time-series analysis in the sense that it allows description of non-linear, multivariate dynamic systems with complex conditional independence structures. Further, the scheme provides a method for efficient backward smoothing and possibilities for efficient, approximate forecasting methods. The scheme has been implemented on top of the HUGIN shell.