Re: st: Efficient way to create adjacency matrix in Mata

["Lacy,Michael" <Michael.Lacy@colostate.edu>]

On Wed, Jun 29, 2011 at 10:01 PM, Benjamin Allaire wrote:
> Question: what is the most efficient way to create an adjacency matrix in mata,
> given a unique identifier ("id") and a club id ("club_id") for folks in the same club.
>
> So the data looks like this:
>
> id    club_id
> 1       1
> 2       1
> 3       2
> 4       2
> 5       2
>
> Is there a clever way to do it without looping over the people?

Here's another approach:

clear
// create example data
set seed 12579
set obs 10
local nclub = 3 
gen int id = _n
gen club_id = 1 + trunc(`nclub' * runiform())
//
// Restructure data so that observations are clubs, variables are member ids w/in clubs
sort club_id
by club_id: gen int nmem = _N if _n == 1
quiet {
	summ nmem
	forval i = 1/`=r(max)' {
   	by club_id: gen mem`i' = id[`i'] if (_n ==1) 
	}
}	
//
// Make some stuff nicer for Mata
local nsubj = _N 
keep if !missing(nmem) 
keep nmem mem*
//
// Mata loops over all non-duplicate pairs within club
mata
  mata clear
  M = J(0,0,.)
  st_view(M, ., .)
  adj = J(`nsubj', `nsubj', 0)
  for (nc = 1; nc <= rows(M); nc++) {     // each club
     nmem = M[nc,1]
     for (i = 2; i <= nmem; i++) {       // all non-duplicating pairs of members w/in club
        for (j = i + 1; j <= nmem + 1 ; j++) {     
           ego = trunc(M[nc,i])
           alter = trunc(M[nc,j])
           adj[ego, alter] = 1
           adj[alter,ego] =1
        }
     }
  }
end


Regards,


Mike Lacy
Assoc. Prof./Dir. Grad. Studies
Dept. of Sociology
Colorado State University
Fort Collins CO 80523-1784
970.491.6721 (voice)

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/