Dynamic Bayesian Networks with Conditional Dynamics in Edge Addition and Deletion

TL;DR

This paper introduces a dynamic Bayesian network framework with conditional dynamics for edge addition and deletion, enabling more intuitive and frequent network evolution, estimated via MCMC and demonstrated in portfolio selection.

Contribution

It proposes a novel dynamic Bayesian network model that allows for more natural edge changes, overcoming limitations of previous mixture network and structural prior approaches.

Findings

Model induces more frequent and intuitive edge changes

Effective in portfolio selection application

Uses MCMC for structure and parameter estimation

Abstract

This study presents a dynamic Bayesian network framework that facilitates intuitive gradual edge changes. We use two conditional dynamics to model the edge addition and deletion, and edge selection separately. Unlike previous research that uses a mixture network approach, which restricts the number of possible edge changes, or structural priors to induce gradual changes, which can lead to unclear network evolution, our model induces more frequent and intuitive edge change dynamics. We employ Markov chain Monte Carlo (MCMC) sampling to estimate the model structures and parameters and demonstrate the model's effectiveness in a portfolio selection application.