Mergeomics (version 1.0.0)

tool.subgraph.stats: Calculate node degrees and strengths

Description

tool.subgraph.stats graph statistics (degrees and strengths) of the seed nodes obtained from their neighborhoods.

Usage

tool.subgraph.stats(frame, edgemap, heads, weights)

Arguments

frame

a data frame including following components:

RANK: indices of neighboring nodes (including seeds)
LEVEL: number of edges away from seed 
STRENG: sum of adjacent edge weights within neighborhood
DEGREE: number of adjacent edges within neighborhood

edgemap

list of adjacent edge information for detected neighborhoods of seed nodes. edgemap can belong to either tails or heads.

heads

list of either head (destination) or tail (source) nodes for neighborhoods of the seed nodes

weights

weights of the edges in the entire graph

Value

following components:

RANK

indices of neighboring nodes (including seeds)

LEVEL

number of edges away from seed

STRENG

sum of adjacent edge weights within neighborhood

DEGREE

number of adjacent edges within neighborhood

Examples

data(job_kda_analyze)
depth <- 1
direction <- 0
## Take one or multiple center nodes (seeds) to search the neighborhoods:
## e.g. take the first node in the graph as the seed, find its neighborhood:
center.node = job.kda$graph$nodes[1]
## Convert center node (seed) names to indices:
nodes <- job.kda$graph$nodes
ranks <- match(center.node, nodes)
ranks <- ranks[which(ranks > 0)]
## we already know that rank is 1, since we took the first node in the graph
## as an example:
ranks <- as.integer(ranks) 
## Find edges to adjacent nodes. (both up- and down-stream searches)
visited <- ranks
levels <- 0*ranks
for(i in 1:depth) {
## Find edges to adjacent nodes.
foundT <- tool.subgraph.find(ranks, job.kda$graph$tail2edge, 
job.kda$graph$heads, visited)
foundH <- tool.subgraph.find(ranks, job.kda$graph$head2edge, 
job.kda$graph$tails, visited)        
## Expand neighborhood for the further depths of the neighborhood search
ranks <- unique(c(foundT, foundH))
visited <- c(visited, ranks)
levels <- c(levels, (0*ranks + i)) ## level shows the depth
if(length(ranks) < 1) break
}
## Calculate node degrees and strengths.
res <- data.frame(RANK=visited, LEVEL=levels, DEGREE=0,
STRENG=0.0, stringsAsFactors=FALSE)
res <- tool.subgraph.stats(res, job.kda$graph$tail2edge, 
job.kda$graph$heads, job.kda$graph$weights)
res <- tool.subgraph.stats(res, job.kda$graph$head2edge,
job.kda$graph$tails, job.kda$graph$weights)