[igraph] Recommend Package for Bonacich Power

[Surendar Swaminathan]

Hello All,

 

  There was Solve Matrix Package error when trying to do Bonaicih power for large graph.  R version 2.8.1 Can someone recommend a package to work on Philip bonacich  power.

 

I posted the same long time back and was directed to the code.

 

http://igraph.wikidot.com/r-recipes#toc6  this is the output I had got

 

solve() errors and crashes

 

I posted the error on R user list and was told to use different code still I could not solve the problem

##
### "Enhanced" by Martin Maechler:

bonpow.sparse <- function(graph, nodes = V(graph), loops = FALSE,

                         exponent = 1, adj.type = "both",
                         trace = TRUE,
                         rescale=FALSE, tol=1e-07)
{
   stopifnot(require("igraph"),
             require("Matrix"))

   if(trace) {
       c.width <- 30
       C1 <- function(s) cat(sprintf("%-*s .. ", c.width, s))
       C2 <- function() cat("[Ok]\n")
   } else { C1 <- C2 <- function(...) {} }

   ## remove loops if requested
   if (!loops) {
       C1("simplify()ing graph")
       graph <- simplify(graph, remove.multiple=FALSE, remove.loops=TRUE)
       C2()
   }

   ## sparse adjacency matrix
   C1("d <- get.adjacency(., sparse)")
   d <- get.adjacency(graph, type = adj.type, sparse=TRUE); C2()

   if(trace >= 2)
       cat("class(d): ", class(d),"\n")
   if(!is.directed(graph)) {
       ## MM: unfortunately  "igraph" does not return a *symmetric*
       ## --- sparse matrix directly, saving space and time,
       ## so we at least do it now :
       C1("d <- as(d, \"symmetricMatrix\")")
       d <- as(d, "symmetricMatrix"); C2()
   }

   ## sparse identity matrix
   vg <- vcount(graph)
   if(FALSE) { ## "non-sense" :
       C1("spMatrix(.) for Diagonal(vg)")
       id <- spMatrix(vg, vg, i=1:vg, j=1:vg, x = rep(1, vg)); C2()
       C1("    --> as(., \"dgCMatrix\")")
       id <- as(id, "dgCMatrix"); C2()
   }
   else
       id <- Diagonal(vg)

   C1("M <- (id - exponent * d)")
   M <- id - exponent * d ; C2()

   C1("b <- degree(graph,.)")
   b <- degree(graph, mode="out") ; C2()

   if(trace >= 2) {
       cat("  M : class ", class(M),";    dim: ", dim(M),"\n")
       cat("  b : class ", class(b),"; length: ", length(b),"\n")
   }

   ## solve it

   ## MM: This is "the horror"  -- ("it's the economy, stupid !!")
   ##     particularly as  solve(M) is *never* sparse !!!
   ## He should use  solve(M, b) !!
   if(FALSE) {
       C1("solve(M) %*% b")
       ev <- solve(M, tol=tol) %*% b; C2()
   } else {
       C1("solve(M, b)")
       ev <- solve(M, b) ; C2()

   }

   if (rescale) {
       ev <- ev/sum(ev)
   } else {

       ev <- ev * sqrt(vg/sum((ev)^2))

- Hide quoted text -
   }

   ev[as.numeric(nodes) + 1]

 

 

 

Thanks in Advance.