turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Nested random effects in a Hurdle model

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

08-25-2009 03:41 PM

Hello,

I have read some papers about including a random effect in a hurdle model, but none about including nested random effects, and would like to know if this is possible.

I have made census in 4 transects, with 20 fields inside each transect. I think using the glimmix procedure the code for the random effects would be:

random transects fields(transects)

But I don't know how to translate this into the hurdle model with the NLMIXED procedure. This is the nearest code I have for this (with only one random effect):

proc nlmixed data=sasuser.hydra_glm_enteros;

parms a0=0 a1=0

b0=0 b1=0

s=1;

eta0 = a0 + a1 * estad_a + u1;

exp_eta0 = exp(eta0);

p0 = exp_eta0 / (1 + exp_eta0);

etap = b0 + b1 * estad_a + u1;

exp_etap = exp(etap);

if rec_avo = 0 then ll = log(p0);

else ll = log(1 - p0) - exp_etap + ard_cin * etap - lgamma(ard_cin + 1)

- log(1 - exp(-exp_etap));

model rec_avo ~ general(ll);

random u1 ~ normal(0,s) subject= transect;

run;

I would appreciate very much any answer.

Thanks,

Grego.

I have read some papers about including a random effect in a hurdle model, but none about including nested random effects, and would like to know if this is possible.

I have made census in 4 transects, with 20 fields inside each transect. I think using the glimmix procedure the code for the random effects would be:

random transects fields(transects)

But I don't know how to translate this into the hurdle model with the NLMIXED procedure. This is the nearest code I have for this (with only one random effect):

proc nlmixed data=sasuser.hydra_glm_enteros;

parms a0=0 a1=0

b0=0 b1=0

s=1;

eta0 = a0 + a1 * estad_a + u1;

exp_eta0 = exp(eta0);

p0 = exp_eta0 / (1 + exp_eta0);

etap = b0 + b1 * estad_a + u1;

exp_etap = exp(etap);

if rec_avo = 0 then ll = log(p0);

else ll = log(1 - p0) - exp_etap + ard_cin * etap - lgamma(ard_cin + 1)

- log(1 - exp(-exp_etap));

model rec_avo ~ general(ll);

random u1 ~ normal(0,s) subject= transect;

run;

I would appreciate very much any answer.

Thanks,

Grego.