epinet (version 2.1.11)

BuildX: Build a dyadic covariate matrix (X)

Description

Build a dyadic covariate matrix (X) from a given nodal covariate matrix.

Usage

BuildX(nodecov, unaryCol = NULL, unaryFunc = NULL, 
	binaryCol = NULL, binaryFunc = NULL, includeIntercept = TRUE)

Value

A dyadic covariate matrix with \({N \choose 2}\) rows, columns 1 and 2 are node ids, column 3 is all ones (if requested) and then one column for each given element of unaryCol and binaryCol.

Assigns colnames depending on type of unaryFunc and binaryFunc and colnames of nodecov.

Arguments

nodecov

an N x k matrix where N is the number of nodes, column 1 is the node id and columns 2:k are covariate values for the node

unaryCol

a vector of column indices

unaryFunc

a vector of the same length as unaryCol of method names for comparing dyads. Possible method names are "match" and "absdiff"

binaryCol

a list of 2 element vectors of column indices

binaryFunc

a vector of the same length as binaryCol of method names for comparing dyads. Possible method names are "euclidean" and "manhattan"

includeIntercept

logical. If TRUE, includes a column of all ones. Defaults to TRUE

Author

David Welch , Chris Groendyke

See Also

SimulateDyadicLinearERGM for simulating a contact network based on a dyadic covairate matrix, and epinet for performing inference on the network and epidemic model parameters.

Examples

# make some nodal covariates
set.seed(3)
mycov = data.frame(id = 1:5, xpos = rnorm(5), ypos = rnorm(5), 
	house = c(1, 1, 2, 2, 2), gender = c(0, 0, 0, 1, 1))
# make matrix 
dyadCov = BuildX(mycov, unaryCol = c(4, 5), unaryFunc = c("match", "match"), 
	binaryCol = list(c(2, 3)), binaryFunc = "euclidean")