Re: st: Adjacency matrix in Stata

[Brendan Halpin <brendan.halpin@ul.ie>]

I wouldn't use Stata matrices here. You can use Stata data sets instead,
with very little limit as regards number of cases, if you structure your
data as (person_i, person_j, X), where X is a variable representing the
strength of the link (e.g., the count of times the same product is
bought on the same day).

Start by creating a data set at individual level, with key variables
product and date. Duplicate it, changing the name of the person-ID
variable in the second. Then do a many-to-many merge using the
product and date as the key, so that you end up with a person_i/person_j pair for
every time they bought the same product on the same day. Then use, for
instance, -collapse- to sum the number of common purchases, delete cases
below your threshold, delete the cases here individuals are matched to
themselves, and you have your adjacency matrix. It's in I/J/X format,
rather than the very sparse square matrix, but the all information is
there.

I don't know what you want to do with it, but network analysis commonly
works with adjacency matrices represented like this, because it is
efficient.

Brendan


Example (untested), if purchases.dta has variables id, product and date:

use purchases
rename id idj
merge m:m product date using purchases

drop if id==idj
sort id idj
gen n = 1
collapse (sum) n, by(id idj)
keep if n>=10



-- 
Brendan Halpin, Head, Department of Sociology, University of Limerick, Ireland
Tel: w +353-61-213147  f +353-61-202569  h +353-61-338562;  Room F1-002 x 3147
mailto:brendan.halpin@ul.ie    ULSociology on Facebook: http://on.fb.me/fjIK9t
http://teaching.sociology.ul.ie/bhalpin/wordpress         twitter:@ULSociology
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/