Many thanks to everyone for the huge help on the outer products problem.
I've run the various solutions on some of my data. R.E. Boss's is the clear
winner. I've also included timings from some Matlab code for comparison. I
did notice the timings fluctuating a bit and I suspect Raul's, Roger's, and
Oleg's versions might converge on roughly the same time. Unfortunately,
while Brian's approach gave comparable timing results to that of R.E. Boss,
it doesn't scale to the size of vector I'm using.
$T
61683
W
276
NB. R.E. Boss
ts 'W REB T'
2.33526 1.07454e9
NB. Matlab
Elapsed time:
60.2370
NB. Raul Miller
ts 'W RM T'
107.119 3.67904e6
NB. Roger Hui
ts '(1+61683-276) %~ > ((*/|.)@[ + ])&.> / (276 <\ T),<0'
110.137 2.57786e8
NB. Oleg Kobchenko
ts 'W OK T'
113.35 5.36872e8
NB. Ronan Reilly
ts 'W RR T'
132.914 3.6768e6
NB. Brian Schott
ts 'W BS T'
|limit error: BS
| (,~x)$(/:@;@([:</.i.)@,~x){;<@(%&(1+x-~#y))@((1+x-~#y)+/\])/. (*"0
1|.)y
--
Professor Ronan Reilly
Department of Computer Science
NUI Maynooth
Maynooth
Co. Kildare
IRELAND
t: +353-1-7083846
e: [EMAIL PROTECTED]
w: http://www.cs.nuim.ie; http://cortex.cs.nuim.ie
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
