multi_gp usage?

[Mike Lawrence]

I just stumbled across the multi_gp & multi_gp_cholesky functions in the manual and wonder if anyone has any examples using these (the GP section of the manual instead uses multi_normal and multi_normal_cholesky). 

I'm particularly trying to figure out how the "w" parameter to multi_gp/multi_gp_cholesky maps to the eta/rho/sigma parameters in the examples using multi_normal. My guess is that if you construct Sigma without multiplying by eta but instead dividing sigma by eta on the diagonal, then w would then map as 1/eta, but possibly I'm incorrect?

Mike

--
Mike Lawrence
Graduate Student
Department of Psychology & Neuroscience
Dalhousie University

~ Certainty is (possibly) folly ~

[Bob Carpenter]

Did you ever figure this out? Feel free to ping the list
again in a couple days if we don't answer.

- Bob

> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.
> To post to this group, send email to stan-...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Mike Lawrence]

No, I don't believe I've figured this out yet, so ... ping! :O)

--
Mike Lawrence
Graduate Student
Department of Psychology & Neuroscience
Dalhousie University

~ Certainty is (possibly) folly ~

[Mike Lawrence]

Someone just emailed me asking if I ever figured this out, which I haven't. 

Daniel Lee, I see from the github commits that you're the most involved in the creation/maintenance of multi_gp. Might you have any examples for its use?

On Thu, Apr 14, 2016 at 5:15 PM, Mike Lawrence <mike....@gmail.com> wrote:

No, I don't believe I've figured this out yet, so ... ping! :O)

--
Mike Lawrence
Graduate Student
Department of Psychology & Neuroscience
Dalhousie University

~ Certainty is (possibly) folly ~

On Thu, Apr 7, 2016 at 7:08 PM, Bob Carpenter <ca...@alias-i.com> wrote:

Did you ever figure this out?  Feel free to ping the list
again in a couple days if we don't answer.

- Bob

> On Jan 26, 2016, at 8:08 PM, Mike Lawrence <mike....@gmail.com> wrote:
>
> I just stumbled across the multi_gp & multi_gp_cholesky functions in the manual and wonder if anyone has any examples using these (the GP section of the manual instead uses multi_normal and multi_normal_cholesky).
>
> I'm particularly trying to figure out how the "w" parameter to multi_gp/multi_gp_cholesky maps to the eta/rho/sigma parameters in the examples using multi_normal. My guess is that if you construct Sigma without multiplying by eta but instead dividing sigma by eta on the diagonal, then w would then map as 1/eta, but possibly I'm incorrect?
>
> Mike
>
> --
> Mike Lawrence
> Graduate Student
> Department of Psychology & Neuroscience
> Dalhousie University
>
> ~ Certainty is (possibly) folly ~
>

> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+unsubscribe@googlegroups.com.

> To post to this group, send email to stan-...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.

To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+unsubscribe@googlegroups.com.

[Daniel Lee]

Sorry, but I didn't write it. I've moved it around, which is why my name's attached to the code, but it's not me.

Perhaps the C++ doc helps?

https://github.com/stan-dev/math/blob/1f2794953e5199717afd87bf40dbbe1f7165ce20/stan/math/prim/mat/prob/multi_gp_log.hpp

There are some tests with concrete examples, but I think I might have put them in just to exercise the code once because I also didn't know what it did:

https://github.com/stan-dev/math/search?utf8=%E2%9C%93&q=multi_gp_log&type=Code

Daniel

[Bob Carpenter]

I don't know how I missed that one going in. I didn't even know we had
such a distribution. It's very specialized. Who did add it? And are
there example models somewhere?

- Bob

> > To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.

> > To post to this group, send email to stan-...@googlegroups.com.
> > For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.

> To post to this group, send email to stan-...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>
>
>
>
> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+...@googlegroups.com.

[Daniel Lee]

This was before the Stan / Math split and was added by Marcus: https://github.com/stan-dev/stan/pull/350

None of the example-models have multi_gp. I don't know of any other examples out there.

Daniel

> > To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+unsubscribe@googlegroups.com.

> > To post to this group, send email to stan-...@googlegroups.com.
> > For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+unsubscribe@googlegroups.com.

> To post to this group, send email to stan-...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>
>
>
>
> --
> You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+unsubscribe@googlegroups.com.

> To post to this group, send email to stan-...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "Stan users mailing list" group.

To unsubscribe from this group and stop receiving emails from it, send an email to stan-users+unsubscribe@googlegroups.com.