nblag {spdep}

Higher order neighbours lists

Description

The function creates higher order neighbour lists, where higher order neighbours are only lags links from each other on the graph described by the input neighbours list. It will refuse to lag neighbours lists with the attribute self.included set to TRUE. nblag_cumul cumulates neighbour lists to a single neighbour list (“nb” object).

Usage

nblag(neighbours, maxlag)
nblag_cumul(nblags)

Arguments

neighbours

input neighbours list of class nb

maxlag

the maximum lag to be constructed

nblags

a list of neighbour lists as output by nblag

Value

returns a list of lagged neighbours lists each with class nb

Author(s)

Roger Bivand [e-mail] and Giovanni Millo

See Also

summary.nb

Examples

columbus <- st_read(system.file("shapes/columbus.gpkg", package="spData")[1], quiet=TRUE)
col.gal.nb <- read.gal(system.file("weights/columbus.gal", package="spData")[1])
coords <- st_coordinates(st_centroid(columbus))
summary(col.gal.nb, coords)
col.lags <- nblag(col.gal.nb, 2)
print(col.lags)
summary(col.lags[[2]], coords)
plot(st_geometry(columbus), border="grey")
plot(col.gal.nb, coords, add=TRUE)
title(main="GAL order 1 (black) and 2 (red) links")
plot(col.lags[[2]], coords, add=TRUE, col="red", lty=2)
cuml <- nblag_cumul(col.lags)
cuml
run <- FALSE
if (require(igraph, quietly=TRUE) && require(spatialreg, quietly=TRUE)) run <- TRUE
if (run) {
W <- as(nb2listw(col.gal.nb), "CsparseMatrix")
G <- graph.adjacency(W, mode="directed", weight="W")
D <- diameter(G)
nbs <- nblag(col.gal.nb, maxlag=D)
n <- length(col.gal.nb)
lmat <- lapply(nbs, nb2mat, style="B", zero.policy=TRUE)
mat <- matrix(0, n, n)
for (i in seq(along=lmat)) mat = mat + i*lmat[[i]]
G2 <- shortest.paths(G)
print(all.equal(G2, mat, check.attributes=FALSE))
}