Our speaker this time will be Reza Mohammadi who will discuss his work with Ernst Wit on the BDgraph package, which is a statistical tool for Bayesian model determination in undirected Gaussian graphical models. During the talk, he will cover the statistical methodology used by the package and then discuss its application to several real and simulation examples.
As usual we'll meet in the cafe from 18:30 with the speaker starting at 19:00.
Graphs are one possible way to model relationships between various actors, i.e. the nodes, in many complex systems. If these actors give rise to noisy data, as is the case in many data-intensive sciences, such as biology, finance and epidemiology, then graphical models present an appealing and insightful way to describe graph-based dependencies between the noisy data. Our aim in this talk is to describe a method for inferring the underlying unknown graph structure from this noisy data.
In the first part of the talk, we briefly present statistical methodology behind BDgraph package, which we recently developed for model selection in graphical models. Then, in second part, by using BDgraph package we show the simplicity and efficiency of the proposed methodology on several real data sets in different sciences and some simulation examples for showing accuracy.
We look forward to seeing everyone and hope you enjoy this interesting talk.
Reza, Øyvind and Chris