SCM

[#1734] scale of start u?

Date:
2012-01-09 22:47
Priority:
3
State:
Open
Submitted by:
Mollie Brooks (mbrooks)
Assigned to:
Nobody (None)
Hardware:
None
Product:
None
Operating System:
None
Component:
None
Version:
None
Severity:
None
URL:
Summary:
scale of start u?

Detailed description
I'm interested in initializing my random effects. I see that I should specify u in the list for the start argument. What scale is u on? If it's not on the same scale as the fixed effects coefficients, it might be helpful to show an example of how to transform coefficients to the correct scale.

It would help me personally to be able to extract all coefficients from a glmer fit and use them as starting values for glmmadmb. I know the starting values shouldn't matter this much, but I'm doing poisson regression with an identity link and glmmadmb keeps starting with function value nan. I am able to fit the exact same model in glmer (see code). I'm wondering if it's a problem with the newly implemented identity link. By the way, thanks for implementing that.

Comments:

No Comments Have Been Posted

Attached Files:

Attachments:
Size Name Date By Download
1 KiBsimple simulate.R2012-01-09 22:47mbrookssimple simulate.R

Changes

Field Old Value Date By
File Added128: simple simulate.R2012-01-09 22:47mbrooks
Thanks to:
Vienna University of Economics and Business Powered By FusionForge