spdep (version 0.8-1)

mstree: Find the minimal spanning tree

Description

The minimal spanning tree is a connected graph with n nodes and n-1 edges. This is a smaller class of possible partitions of a graph by pruning edges with high dissimilarity. If one edge is removed, the graph is partioned in two unconnected subgraphs. This function implements the algorithm due to Prim (1987).

Usage

mstree(nbw, ini = NULL)

Arguments

nbw

An object of listw class returned by nb2listw function. See this help for details.

ini

The initial node in the minimal spanning tree.

Value

A matrix with n-1 rows and tree columns. Each row is two nodes and the cost, i. e. the edge and it cost.

Details

The minimum spanning tree algorithm.

Input a connected graph.

Begin a empty set of nodes.

Add an arbitrary note in this set.

While are nodes not in the set, find a minimum cost edge connecting a node in the set and a node out of the set and add this node in the set.

The set of edges is a minimum spanning tree.

References

R. C. Prim (1957) Shortest connection networks and some generalisations. In: Bell System Technical Journal, 36, pp. 1389-1401

Examples

# NOT RUN {
### loading data
require(maptools)
bh <- readShapePoly(system.file("etc/shapes/bhicv.shp",
      package="spdep")[1])
### data padronized
dpad <- data.frame(scale(bh@data[,5:8]))

### neighboorhod list 
bh.nb <- poly2nb(bh)

### calculing costs
lcosts <- nbcosts(bh.nb, dpad)

### making listw
nb.w <- nb2listw(bh.nb, lcosts, style="B")

### find a minimum spanning tree
system.time(mst.bh <- mstree(nb.w,5))

dim(mst.bh)

head(mst.bh)
tail(mst.bh)

### the mstree plot
par(mar=c(0,0,0,0))
plot(mst.bh, coordinates(bh), col=2, 
     cex.lab=.7, cex.circles=0.035, fg="blue")
plot(bh, border=gray(.5), add=TRUE)

# }