[ADMB Users] constrained correlation matrices

[dave fournier]

On 12-11-23 02:43 AM, H. Skaug wrote:


For the correlation matrix parameterization  think you should start with 
k=1 or
use ++k instead of k++.
> I have included the zero-constraint as part of a description of how
> to parameterize covariance matrices.
>
> http://www.admb-project.org/examples/admb-tricks/parameterization/covariance-matrices
>
> Hans
>
> On Thu, Nov 22, 2012 at 6:05 PM, dave fournier <davef at otter-rsch.com> wrote:
>> Well I came across a paper  that was dealing with this problem but I can't
>> find it now.
>> More interesting is to extend the method to  constraints like
>>
>>      rho_ij = .7
>>
>> which I think can be done so long as the constraints are consistent.
>>
>> As another approach I think that more or less arbitrary consistent
>> constraints
>> can be implemented using augmented Lagrangian methods.
>>
>> _______________________________________________
>> Users mailing list
>> Users at admb-project.org
>> http://lists.admb-project.org/mailman/listinfo/users