Vignettes -- proposal

[BERENZ]

I propose the following vignettes:

1. Models -- describing distributions and models
2. Estimation of the population size -- a document that describes how the pop size and its variance is estimated
3. Diagnostics -- post-hoc diagnostics (i.e. LOO and others)
4. Computation -- document covering estimation procedure (theory + parameters)
5. Examples -- reproducing results from selected papers and some tutorials
6. Comparison with other R packages
7. Literature

[BERENZ]

I can take nr 5 and 6.

[Kertoo]

• 1. is already done in help file that can be accessed from references https://ncn-foreigners.github.io/singleRcapture/reference/singleRmodels.html
• 2. is similarly already done in https://ncn-foreigners.github.io/singleRcapture/reference/estimatePopsize.html
• 4. is done in https://ncn-foreigners.github.io/singleRcapture/reference/estimatePopsize.html and https://ncn-foreigners.github.io/singleRcapture/reference/estimatePopsize.fit.html
• 5. is done as most help files have examples
• 7. is done as https://ncn-foreigners.github.io/singleRcapture/reference/estimatePopsize.html has extensive citations
• 6. has some work done with comparing singleRcapture with VGAM
• 3. is unnecessary since all diagnostics are just from stats::glm and VGAM::vglm and sandiwch::vcovHC etc. The only one unique to singleRcapture is dfpopsize which already has a help file (which admittedly can be more specific)

[Kertoo]

I think it would be more usefull to just add get started page on ncn-foreigners.github.io/singleRcapture with some basic tutorial similar to e.g https://pkgdown.r-lib.org/articles/pkgdown.html

[BERENZ]

I think it would be nice to have vignettes that organize the knowledge. For instance:

• Models -- describing distributions and models -- it could be a list of models with reference to specific help function with a description what is the relation between them etc.
• Examples -- reproducing results from selected papers and some tutorials -- this can contain reference to the publications and showing that our package reproduce these results
• Comparison with other R packages -- reference to VGAM, countreg and others to show superiority of this package for SSCR.

[Kertoo]

Okay that makes sense so can you manage reproduction of results from existing papers with maybe some examples of how we can use additional features in singleRcapture to use more accurate model/test whether model is correct i.e. no one inflation by e.g. lmtest::lrtest or adding formula for omega/alpha parameters and comparisons with countreg and VGAM? I'll handle models with their interpretation and some note on computational quirks of package functions with maybe some updated tutorial.