network.scan: Network Scan Statistic

In SNscan: Scan Statistics in Social Networks

Description

Evaluate scan statistics in social network.

Usage

network.scan(g, radius, attribute, model, pattern, 
	max.prop = 0.5, xmin = NULL, zetatable = NULL) 

Arguments

g	An igraph object.
radius	The radius of scanning windows. Default is 3.
attribute	The interested attribute which should be a data list including observations (obs) and populations (pop).
model	The distribution of attribute which can be "norm.stat", "pois.stat", "binom.stat", "multinom.stat", and "powerlaw.stat".
pattern	The testing pattern of the network which can be "structure", "attribute", and "both".
max.prop	Numeric value, the maximum proportion of selecting graph. Default is 0.5.
xmin	Numeric value, the minimum value only for powerlaw stat. Default is 1.
zetatable	Zatatable is applied when power-law distribution is used. Default is NULL.

Value

A matrix will be returned. The values include C: the center of a scanning window, D: The radius of a scanning window. test.L: the test statistic, S0: Indicating the testing information within the selected nodes. Sz: Indicating the testing information outside the nodes. The final column, z.length, is the number of the nodes in the cluster with corresponding center and radius.

Author(s)

Taichi Wang <[e-mail]>

See Also

network.mc.scan

Examples

data(karate)

ks=network.scan(g=karate,radius=3,attribute=NULL,
	model="pois.stat",pattern="structure")	
mc.ks=network.mc.scan(n=9,g=karate,radius=3,attribute=NULL,
	model="pois.stat",pattern="structure")	
pv=mcpv.function(obs.stat=ks[,3],ms.stat=mc.ks[,3],direction=">=")

SNscan documentation built on May 2, 2019, 8:26 a.m.