networktools (version 1.1.0)

structure.impact: Network Structure Impact

Description

Generates the network structure impact of each specified node. Network structure impact can be interpreted as the degree to which the level of a node causes change in the network structure

Usage

structure.impact(input, gamma, nodes = c("all"), binary.data = FALSE,
  weighted = TRUE, split = c("median", "mean", "forceEqual", "cutEqual",
  "quartiles"))

Arguments

input

a matrix or data frame of observations (not a network/edgelist). See included example datasets depression and social.

gamma

the sparsity parameter used in generating networks. Defaults to 0.5 for interval data and 0.25 for binary data

nodes

indicates which nodes should be tested. Can be given as a character string of desired nodes (e.g., c("node1","node2")) or as a numeric vector of column numbers (e.g., c(1,2)).

binary.data

logical. Indicates whether the input data is binary

weighted

logical. Indicates whether resultant networks preserve edge weights or binarize edges. Note: unweighted networks will always result in a network structure impact of 0 or 1.

split

method by which to split network given non-binary data. "median": median split (excluding the median), "mean": mean split, "forceEqual": creates equally sized groups by partitioning random median observations to the smaller group, "cutEqual": creates equally sized groups by deleting random values from the bigger group,"quartile": uses the top and bottom quartile as groups

Value

structure.impact() returns a list of class "structure.impact" which contains:

impact

a named vector containing the network structure impact for each node tested. Network structure impacts are given as absolute values

edge

a list of vectors. Each vector contains a the edge impact of the most strongly impacted edge (e.g., the network structure impact)

lo

a named vector containing the edge estimate for the lower half of the most strongly impacted edge

hi

a named vector containing the edge estimate for the upper half of the most strongly impacted edge

Details

For an explanation of impact functions in general, see impact.

Network structure impact computes network structure invariance as a function of node level. Network structure invariance is defined as the absolute value of the single largest edge invariance between two networks. A large network structure invariance is generally used to indicate the instability of network structure across groups. For instance, a low global strength invariance coupled with a high network structure invariance would mean that although the overall connectivity remains stable, the actual structure of those edges is unstable across groups.

Examples

# NOT RUN {
out <- structure.impact(depression[,1:3])
# }
# NOT RUN {
out1 <- structure.impact(depression)
out2 <- structure.impact(depression, gamma=0.65,
    nodes=c("sleep_disturbance", "psychomotor_retardation"))
out3 <- structure.impact(social, binary.data=TRUE)
out4 <- structure.impact(social, nodes=c(1:6, 9), binary.data=TRUE)

summary(out1)
plot(out1)

#Determine which edge drove network structure impact of "sadness"
out1$edge$sadness
# }