simulation_instance_ER: Shuffling Simulation for ER graphs
In dpmcsuss/gmmle:

Run a simulation instance of different graph models by shuffling vertices and computing disagreement statistics. Models include Erdos Renyi (ER), Watts-Strogatz small world graphs with rewiring probabilities of 0.05 and 0.7, and preferential attachment (PA) models with powers of 1, 1, and 2, and zero.appeal of 0, 500, and 0. All graphs are undirected with n=500 vertices. Each graph is sampled once and the shuffles are repeated m times.

simulation_instance_ER(
  k_grid = c(2, 10, 50, 100, 200, 400),
  p_vec = (1:5)/10,
  n = 500,
  m = 1000,
  seed = 0,
  ask = TRUE
)

`k_grid`	sequence of values for number of shuffled vertices
`p_vec`	vectors of edge probabilities
`n`	number of nodes
`m`	number of Monte Carlo replicates
`seed`	Seed for random number generator
`ask`	Ask about running big simulations.

Tibble with rows for each graph-K pair and 7 columns, one for the model, parameters (list column), and K. Columns 4:6 give the mean, sd, and min for the number of edge disagreements. Column 7 gives a theoretical estimate for the maximum probability of edge flips for which graph matching will still be possible. simulation_instance_ER(k_grid = c(2, 10), p_vec = (1:2)/10, n = 30, m = 10, seed = 0, ask = FALSE)

dpmcsuss/gmmle documentation built on July 2, 2020, 6:24 p.m.