gs.siem.fit: Structured Independent Edge Model, Single Sample

In neurodata/graphstats: Graph Statistics

Description

An independent-edge generalization of the traditional Stochastic Block Model.

Usage

gs.siem.fit(graph, C, alt = "greater", ...)

Arguments

graph	[v, v] the binary graph with v vertices.
C	[[r]] a list of r communities where each element corresponds to a community of edges. Note that union(C[[j]]) over all j communities should be 1:v, and intersect(C[[j]]) over all j communities should be empty.
alt	the alternative hypothesis for each edge-community test. Defaults to 'greater'. 'greater'p corresponds to H0: x1 <= x2, HA: x1 > x2. R(x1, x2) = x1 > x2 'neq'p corresponds to H0: x1 = x2, HA: x1 != x2. R(x1, x2) = x1 != x2 'less'p corresponds to H0: x1 >= x2, HA: x1 < x2. R(x1, x2) = x1 < x2
...	trailing args

Value

An object of class "SIEM" containing the following:

pr	[r] a probability vector corresponding to the probability of an edge connecting each edge community.
var	[r] the variance of the probabilities estimated between edge communities.
dpr	[r, r] an array consisting of the paired differences in probability where dpr_ij = pr_i - pr_j.
dvar	[r, r] an array consisting of the variance of the estimate of the paired differences in probability, where dvar_ij = var_i + var_j.
pv	[r, r] pv_ij is the p-value of false rejection of H0 that !R(pr_i, pr_j) in favor of HA that R(pr_j, pr_i).

Author(s)

Eric Bridgeford

neurodata/graphstats documentation built on May 14, 2019, 5:19 p.m.