Given a series of networks,
prod.network attempts to form a new network by multiplication of edges. If a non-null
attrname is given, the corresponding edge attribute is used to determine and store edge values.
one or more
the name of an edge attribute to use when assessing edge values, if desired.
logical; should edges with missing data be ignored?
The network product method attempts to combine its arguments by edgewise multiplication (not composition) of their respective adjacency matrices; thus, this method is only applicable for networks whose adjacency coercion is well-behaved. Multiplication is effectively boolean unless
attrname is specified, in which case this is used to assess edge values – net values of 0 will result in removal of the underlying edge.
Other network attributes in the return value are carried over from the first element in the list, so some persistence is possible (unlike the multiplication operator). Note that it is sometimes possible to “multiply” networks and raw adjacency matrices using this routine (if all dimensions are correct), but more exotic combinations may result in regrettably exciting behavior.
Carter T. Butts email@example.com
Butts, C. T. (2008). “network: a Package for Managing Relational Data in R.” Journal of Statistical Software, 24(2). http://www.jstatsoft.org/v24/i02/
1 2 3 4 5 6 7 8 9 10 11 12
#Create some networks g<-network.initialize(5) h<-network.initialize(5) i<-network.initialize(5) g[1:3,,names.eval="marsupial",add.edges=TRUE]<-1 h[1:2,,names.eval="marsupial",add.edges=TRUE]<-2 i[1,,names.eval="marsupial",add.edges=TRUE]<-3 #Combine by addition pouch<-prod(g,h,i,attrname="marsupial") pouch[,] #Edge values in the pouch? as.sociomatrix(pouch,attrname="marsupial") #Recover the marsupial