Subject: What's the best practice to specify df for TPrior and alphaByLambda for ElasticBandPrior


This paper has good guidance.

I think that their recommendation of df=1 and scale=2.5 is very reasonable.
 Their suggestion of mean subtraction would be disastrous, but much of the
rest is very good.

df=0.5 is a very (VERY) long tailed distribution.  Even for df=1, there is
no mean or variance.

I would actually consider using df=2 as well.

In practice, I have found that the L1 prior encourages sparsity more
strongly than the T prior.  It does, however, entangle the regularization
factor with the scale of the prior.

On Mon, Jun 6, 2011 at 1:06 AM, XiaoboGu <[EMAIL PROTECTED]> wrote: