bootdht returns data.frame with 0 columns and 0 rows
Milou Groenenberg
Dear group
I am trying out the code to derive the sampling distribution of pairwise differences in estimated density using the bootstrap as in this vignette. Unfortunately, I seem to be having a problem with the “bootdht” function when applying it to my data and model (it works fine using the example datasets of the vignette). My data is for one species and one area across 4 different survey years. If have used the different years as Region.Label to have stratified results. I am using formula = ~size to accommodate the cluster size bias that was clearly observed for this species. My model:BSDPPWS_hn_t81_size <- ds (data=BSDPPWS,
truncation=81,
key="hn",
adjustment=NULL,
convert_units=conversion.factor,
formula=~size)
The problem arises when using the bootdht function, for example:est.bootn <- bootdht(model=BSDPPWS_hn_t81_size,
flatfile=BSDPPWS,
summary_fun=bootdht_Nhat_summarize,
convert_units=conversion.factor,
sample_fraction=1,
nboot=10, cores=1)
*Note: nboot was set low for trying out the code
Est.bootn is then returned as a data.frame with 0 columns and 0 rows
I don't have any problems obtaining the results from my model using the conventional summary() function.
I can see a difference between my model object and those of
the example datasets: my model$dht is a list[3}(S3:dht) and includes individuals,
as well as clusters and Expected.S. The example dataset only include individuals.
Could something be going wrong there when calling the boothdt function?
Finally, for another dataset where estimation was derived via multiple call to ds(), I would also like to compare density differences using bootstrapping. In the limitations sections of the vignette it is stated “However, based upon the provided code, it should be clear how to produce replicate density estimates via bootdht() and then difference them with a single line of code.” Would this be achieved by merging the bootdht() results from the two different models into one outcome dataframe and then running the remaining code as is?
Many thanks in advance for your support!
Milou
Eric Rexstad
ClusterExercise that ships with the Distance package).- what version of the Distance package are you using (version 1.0.7 is the current CRAN version)
- does your BSDPPWS data frame contain a column for area of the strata? If area is not provided, an abundance estimate cannot be produced (only a density estimate). Because you are harvesting the abundance estimates from the bootstrap
via the
bootdht_Nhat_summarize function, if those abundances do not exist, code might return a data.frame with 0 columns and 0 rows). Recognize there is a companion summary functionbootdht_Dhat_summarize for use when inference is based around density estimates.
Sent: 21 February 2023 03:10
To: distance-sampling <distance...@googlegroups.com>
Subject: [distance-sampling] bootdht returns data.frame with 0 columns and 0 rows
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/6d7868e9-023f-411c-9032-50870971921cn%40googlegroups.com.
Milou Groenenberg
Distance package, I mean updating from Github, rather than updating from CRAN. For a Github update, use these two lines of code:Eric Rexstad
Region.Label that begin with letters that appear in the alphabet before
T. We hope to get the bug sorted, but meanwhile follow this advice when naming your strata for analysis.To: Eric Rexstad <Eric.R...@st-andrews.ac.uk>
Cc: distance-sampling <distance...@googlegroups.com>
Subject: Re: [distance-sampling] bootdht returns data.frame with 0 columns and 0 rows