FAQ

FAQ on use of R package 'marked'

FAQ

Postby egc » Fri Oct 03, 2014 1:03 pm

'marked' FAQ

  • What is marked? marked is an R package that was designed to provide models for capture-recapture analysis that are not currently in MARK (e.g., mixed effect models) or to provide some features that are not currently available in MARK (varying time intervals). It also was a tool for us to explore alternative approaches (e.g. ADMB, MCMC) of fitting models to capture-recapture data.

  • What do I need to run marked? - You need a recent version (2.15 or greater) of R and you need to download and install the marked package. A version of marked is available on CRAN and you can download and install it as you would any package. The source code for marked is at

    https://github.com/jlaake/marked

    The code in the master branch is the version on CRAN and the version in the devel branch is the current working version. For Windows user's I keep a binary version of more recent development package at:

    https://drive.google.com/folderview?ddr ... VVGS0YtWE0

    To install the Windows binary, download the file and use the Packages/Install from local zip from the R menu. Simply point to the downloaded file and it will be installed. For non-Windows users, you'll need to install ADMB.

  • What platforms are supported? Any platform that R supports will run marked. On non-Windows machines, you'll also need to install ADMB to use it.

  • Do I need ADMB to run marked? For the most part no, but models can run faster with ADMB. All of the models in marked can be run without ADMB except for the CJS likelihood models with random/mixed effects. The Windows binary mentioned in #2, includes executable files that allow you to use ADMB without having ADMB installed on your computer.

  • Should I use marked? marked is not nearly as mature as MARK and marked currently only contains a few models. If you are new to mark-recapture analysis, I suggest you learn MARK and use RMark for the interface if you want to use R. You'll find the documentation for MARK and RMark is much more complete than what is currently available for marked. Also the software is still early in its development and the interface is still being changed in places and the code is not nearly as well debugged. That said, if you are adventuresome and want to use some of the features and models in marked please do so. Also, feel free to peruse the code and make suggestions. Part of our reasoning for developing marked was to explore alternative analysis approaches (e.g., hidden Markov models) and to provide open-source code that others can use or learn from. Users of RMark will see some similarities with marked and the learning curve will be less.

  • What should I do if I have a problem using marked? Please first read the documentation (help files and vignette) and if you can't figure it out, post a message to this forum. Unless I'm in the field I'll respond within a day or so.

  • What citation should I use for marked?

    Laake, J. L., D. S. Johnson, and P. B. Conn. 2013. marked: An R package for maximum-likelihood and MCMC analysis of capture-recapture data. Methods Ecol. Evol. 4:885–890.

    You can acces the paper online here: http://onlinelibrary.wiley.com/doi/10.1 ... .12065/pdf
egc
Site Admin
 
Posts: 173
Joined: Thu May 15, 2003 3:25 pm

Return to FAQ

Who is online

Users browsing this forum: No registered users and 1 guest