Compute the dissimilarity between 2 mesoscale structures for bipartite SBM

Compute the dissimilarity between 2 mesoscale structures for bipartite SBM

Usage

dist_bisbmpop_max(pi, rho, alpha, delta = c(1, 1), weight = "max", norm = "L2")

Arguments

pi

A list of two probability vectors (row)

rho

A list of two probability vectors (columns)

alpha

A list of two connectivity matrices

delta

A vector of 2 density parameters (optional)

weight

One of "max" (default) or "mean". See details

norm

"L1"or "L2" norm for the computation

Value

The dissimilarity between two mesoscale structure.

Details

If weight is "max" then the weight of each block is computed as pmax(pi[[1]], pi[[2]]). If "mean", then we take the average. "max" penalize to a greater extent the difference in block proportion between structure.

Examples

pi <- list(c(0.5, 0.5), c(0.1, 0.9))
rho <- list(c(0.1, 0.9), c(0.5, 0.5))
alpha <- list(
  matrix(c(
    0.9, 0.1,
    0.1, 0.05
  ), byrow = TRUE, nrow = 2),
  matrix(c(
    0.9, 0.1,
    0.1, 0.05
  ), byrow = TRUE, nrow = 2)
)
dist_bisbmpop_max(pi, rho, alpha)
#> [1] 0