Figure 13
From: Bayesian inference of transition matrices from incomplete graph data with a topological prior
Effect of using different concentration parameters α alongside the topological prior on the powergrid network. Rows indicate the task (top to bottom: transition matrix inference, node ranking, clustering), columns the different values for α (left to right: \(\alpha = 0.01\), \(\alpha = 1\), \(\alpha = 10\)). We observe that compared to the value used in the previous experiment (\(\alpha = 1\)) BaCon performance on the powergrid network improves for the larger α (right column)