On Identifying Significant Edges in Graphical Models
Marco Scutari, Radhakrishnan Nagarajan · arXiv (Cornell University) · 2011
Graphical models, and in particular Bayesian networks, have been widely used to investigate data in the biological and healthcare domains. This can be attributed to the recent explosion of high-throughput data across these domains and the importance of understanding the causal relationships between the variables of interest. However, classic model validation techniques for identifying significant edges rely on the choice of an ad-hoc threshold, which is non-trivial and can have a pronounced impact on the conclusions of the analysis. In this paper, we overcome this limitation by proposing simple, statistically-motivated approach based on L1 approximation for identifying significant edges. The effectiveness of the proposed approach is demonstrated on gene expression data sets across two published experimental studies.