Constrained personalised page rank — constrained_page

This function implements a constrained personalised page-rank. You can define sink_nodes or sink_edges. In the case of the former, the vist of a sink_node (via a type attribute in the igraph) will automatically cause the surfer to reset. In the case of the latter, the traversal of a sink_edge is allowed, however, subsequently, the surfer will be reseted. This function can be useful for deriving PPR profiles under constraints in heterogenous graphs and has been inspired by Ruiz, et al.

Usage

constrained_page_rank(
  graph,
  personalisation_vector,
  sink_nodes = NULL,
  sink_edges = NULL
)

Arguments

graph: igraph. This one needs to be directed and weighted and have the node attribute type defining the node type and the edge attribute type defining the edge type.
personalisation_vector: Numerical vector. The personalisation vector for the reset of the personalised page rank. The function will automatically normalise it.
sink_nodes: Optional String vector. The node types that should force a reset.
sink_edges: Optional String vector. The edge types after which there should be a forced reset.

Value

The constrained personalised page rank

References

Ruiz, et al., Nat Commun, 2021