DiagrammeR (version 0.9.2)

add_pa_graph: Add a preferential attachment graph

Description

To an existing graph object, add a graph built according to the Barabasi-Albert model, which uses preferential attachment in its stochastic algorithm.

Usage

add_pa_graph(graph, n, m = NULL, power = 1, out_dist = NULL,
  use_total_degree = FALSE, zero_appeal = 1, algo = "psumtree")

Arguments

graph

a graph object of class dgr_graph.

n

the number of nodes comprising the preferential attachment graph.

m

the number of edges to add in each time step.

power

the power of the preferential attachment. The default value of 1 indicates a linear preferential attachment.

out_dist

a numeric vector that provides the distribution of the number of edges to add in each time step.

use_total_degree

a logical value (default is TRUE) that governs whether the total degree should be used for calculating the citation probability. If FALSE, the indegree is used.

zero_appeal

a measure of the attractiveness of the nodes with no adjacent edges.

algo

the algorithm to use to generate the graph. The available options are psumtree, psumtree-multiple, and bag. With the psumtree algorithm, a partial prefix-sum tree is used to to create the graph. Any values for power and zero_appeal can be provided and this algorithm never generates multiple edges. The psumtree-multiple algorithm also uses a partial prefix-sum tree but the difference here is that multiple edges are allowed. The bag algorithm places the node IDs into a bag as many times as their in-degree (plus once more). The required number of cited nodes are drawn from the bag with replacement. Multiple edges may be produced using this method (it is not disallowed).

Examples

# NOT RUN {
# Create an undirected PA
# graph with 100 nodes, adding
# 2 edges at every time step
pa_graph <-
  create_graph(
    directed = FALSE) %>%
  add_pa_graph(
    n = 100,
    m = 1)

# Get a count of nodes
pa_graph %>% node_count()
#> [1] 100

# Get a count of edges
pa_graph %>% edge_count()
#> [1] 99
# }