[R-sig-ME] hard in lmer(),how to set parameter in lmer()

Ben Bolker bbolker at gmail.com
Tue Apr 22 03:57:44 CEST 2014

王唯 <wzwei1636 at ...> writes:

>  Dear professor: I apologize for disturbing you! I'm appreciating
> your package 'lme4()' in the R ,because it's very useful when we
> deal with the linear mixed model. However,I get struck in doing my
> Master's degree paper , I'm hoping that you can help me.

  Can you possibly repost your question using only standard ASCII
characters that don't get mangled in translation to the mailing list/
my computer?

>        Y = X*β + B*a + ε   is a linear mixed model , 
  X : dim=n*k .  
  β: dim=k*1 represent fixed effects .
  B :dim=n*(n-2) .  
  a: dim=p*1 represent random effects with a ∼ N(0, τI). 

For example, n=100,k=15 , X and B is known,how can I use lmer() to get
 β and τ .How to set parameters in lmer()?

  You should take a look at the ?modular help page for lme4: you can
set up a bare-bones model and then modify X and Z (corresponding to
X and B in your notation) as you like.

  You might have trouble fitting a reasonable model
with k=15 fixed-effect parameters with only n=100 observations ...

More information about the R-sig-mixed-models mailing list