Some functions for fitting models to distance sampling data using NIMBLE

[Michael Scroggie]

Hi Folks,

I've put together an R package containing some functions for fitting models to distance detection data collected on lines and points using NIMBLE. 

 https://github.com/scrogster/nimbleDistance

The package included some user-defined distributions for NIMBLE that allow fitting of standard half-normal and hazard-rate distance detection functions, along with a brief vignettte describing their use. The functions facilitate the analysis of distance detection data collected on either line transects or points.

The package is at an early stage of development so any suggestions or corrections would be very gratefully accepted. 

Regards,

Michael Scroggie.

[Perry de Valpine]

Thanks Michael!

-Perry

--
You received this message because you are subscribed to the Google Groups "nimble-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nimble-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nimble-users/c435e0ef-8edc-4ef1-8954-d8de1af359c5n%40googlegroups.com.

[Daniel Turek]

Michael, this is fantastic to see.  Thanks for letting us know, and keep in touch any time if you have questions about using NIMBLE.

Nice use of nimbleRcall for the integrator.

And, from looking at nimble_functions.R, it looks like the purpose of the "Rdist" argument isn't entirely clear from our documentation - and now thinking about it, it does seem like a total misnomer calling it "Rdist" when this argument is *not* specifically dealing with the r* random generation function (e.g., rHN, or rHN_V from your package).  Rather, the Rdist argument deals with re-parametizations of the distribution, from the way(s) a user specify parameters to your distribution in a model, to the "canonical" distribution that NIMBLE will use under-the-hood.  But anyway, for your distributions, the "Rdist" argument is not necessary.

Thanks again!  This is great to see.

--

[Michael Scroggie]

Many thanks for the feedback and encouragement Daniel,

Yes, I was a bit confused about the purpose of "Rdist". I'll remove it shortly as I agree that it is unnecessary for my functions.

Thanks for all of your efforts, and those of the rest of the Nimble team. 

Regards,

Michael Scroggie

[Chloe Nater]

Hi Michael!

I am just starting to work with my first distance sampling model (after using nimble for a variety of other model types) and was just looking up whether anyone in the community had already written a function for half-normal distributions. And ta-daa, here it is, and not only that but a whole package! I'll definitely be looking into it. Thanks for doing this, and for posting here that you did :-D