A statistical tools hurdle.

[barking_iguana]

Ugh. R won't do the large multinomial regression I need. It has an addressing limit of 2^32. There are bignum packages to get around that, but they're nto compatible with the multinomial package. Or so says my math guy who's been investigating this.

Anyone know whether SPSS would do the trick or other things I might look at if I'm willing to spend some money/ (Not that I have the money yet.)

Comments

[kent-allard-jr.livejournal.com]

I don't know about SPSS -- I haven't used it in years -- but StataSE 9 won't let me allocate more than1gb.