netdiffuseR (version 1.17.0)

vertex_covariate_compare: Comparisons at dyadic level

Description

Comparisons at dyadic level

Usage

vertex_covariate_compare(graph, X, funname)

Arguments

graph

A matrix of size $n*n$ of class dgCMatrix.

X

A numeric vector of length $n$.

funname

Character scalar. Comparison to make (see details).

Value

A matrix dgCMatrix of size $n*n$ with values in the form of $funname(X[i],X[j])$.

Details

This auxiliary function takes advantage of the sparcity of graph and applies a function in the form of $funname(X[i],X[j])$ only to $(i,j)$ that have no empty entry. In other words, applies a compares elements of X only between vertices that have a link; making nlinks(graph) comparisons instead of looping through $n*n$, which is much faster.

funname can take any of the following values: "distance", "^2" or "quaddistance", ">" or "greater", "<"< code=""> or "smaller", ">=" or "greaterequal", "<="< code=""> or "smallerequal", "==" or "equal".

Examples

# Basic example ------------------------------------------------------------
set.seed(1313)
G <- rgraph_ws(10, 4, .2)
x <- rnorm(10)

vertex_covariate_compare(G, x, "distance")
vertex_covariate_compare(G, x, "^2")
vertex_covariate_compare(G, x, ">=")
vertex_covariate_compare(G, x, "<=")