argument is not a matrix for uniform Key in distance package

[Aditya Ghoshal]

Hi everyone, 

while running uniform models (for all adjustment combinations) in the distance package I sometimes encounter this issue. 

Error in t.default(clusters$vc$detection$partial) : argument is not a matrix

There is no issue for running the other key+adjustments pairs with the same datasets. I just wanted to understand why this is the case.  

Thanks,

[Eric Rexstad]

Aditya

I will log this issue with other software issues in the Distance​ package.

I suspect this message arises when through model selection, the Distance​ package concludes that no adjustment terms are necessary in addition to the uniform key function. In this situation, there are no parameters in the model to be estimated (the uniform key function by itself has no parameters).

As a consequence of fitting a model with no parameters, when the line of code is encountered requesting information about the detection function parameters: clusters$vc$detection$partial, this error arises.

For distance sampling data, it is uncommon for a uniform key without adjustments to be a suitable model for making inference. As a consequence, this message is seldom of great consequence.

---

From: distance...@googlegroups.com <distance...@googlegroups.com> on behalf of Aditya Ghoshal <aditya...@ncf-india.org>
Sent: 30 May 2023 11:51
To: distance-sampling <distance...@googlegroups.com>
Subject: [distance-sampling] argument is not a matrix for uniform Key in distance package

 

--
You received this message because you are subscribed to the Google Groups "distance-sampling" group.
To unsubscribe from this group and stop receiving emails from it, send an email to distance-sampl...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/distance-sampling/41dad7b7-213f-438e-b989-f05f21c6147cn%40googlegroups.com.

[Aditya Ghoshal]

Thank you for the prompt response Eric. I also encounter this issue if I use adjustment=NULL with the uniform key (this particular thing has happened always with the uniform key). So I assumed the uniform key always requires an adjustment term, because without one it is unable to fit the decreasing sightings with distance (due to the basic shape of uniform distributions). 

Regards, 

[Eric Rexstad]

You are correct.

---

From: distance...@googlegroups.com <distance...@googlegroups.com> on behalf of Aditya Ghoshal <aditya...@ncf-india.org>

Sent: 30 May 2023 12:07
To: distance-sampling <distance...@googlegroups.com>
Subject: Re: [distance-sampling] argument is not a matrix for uniform Key in distance package

 

To view this discussion on the web visit https://groups.google.com/d/msgid/distance-sampling/e90a990c-bc03-4dd7-a195-eaf954ceb8b4n%40googlegroups.com.

[Aditya Ghoshal]

Cool. Thank you.