Setting vertex_weights in igraph's implementation of InfoMap

Question

I'm using the community_infomap in igraph to perform community detection in a directed network. My understanding of the algorithm through reading the original paper is that the InfoMap first runs a PageRank like random walk on the network to encode nodes with code words whose lengths are reverse proportional to their rank mass scores towards achieving the final goal of minimizing the description length of the network's topology.

In the igraph's implementation, one can provide both a vector of edge weights and a vector of vertex weights to perform a weighted random walk. It's immediately clear to me of how the edge weights bias the random walk but I'm confused of what a role do the vertex weights play. Does the vertex_weight vector corresponds to (A) the personalized teleportation vector? Or does it correspond to (B) the a priori rank mass scores? Or is it something else?

Thank you very much for your kind answer.

Answer 1

The weights of the nodes correspond to the personalized teleportation vector, judging from the related source files. The weights specified by the user are stored in the teleportWeight member variables of the node objects.