RMark | an alternative approach to DM's in MARK
Posted: Sun Aug 13, 2006 4:01 pmis now available. Thanks to the stellar efforts of Jeff Laake and Eric Rexstad, the RMark library for interfacing R with MARK is now available for public consumption. The library itself can be downloaded directly from this mirror at
http://www.phidot.org/software/mark/download/RMark.zip
or from Jeff's NOAA site.
In addition, Jeff and Eric have recently completed a comprehensive 60+ page 'primer' on using RMark with MARK - it cmprises the new 'Appendix C' for 'the book' - and can be downloaded from the usual sources, or directly here at
http://www.phidot.org/software/mark/docs/book/app_3.pdf
Why should you look at RMark? Well, as most experienced MARK users will confirm, there will come a time when your models are so big, and so complex, that the thought of building a design matrix manually fills you with something approaching dread. You might even end up saying a 'bad word' (or several) if your DM doesn't work. So, in part to simplify your life, RMark was designed to let you build complex models without ever having to 'play' with the design matrix. Using a natural langauge approach to building models within the RMark library, you can quickly build even complex models for most of the available data types currently available in MARK, using a fairly straightforward model building syntax. The use of a natural language approach to building models is not new - it exists to some degree in POPAN - and very much so in M-SURGE. However, the RMark approach also gives you access to all the other functionality of R, which is a definite plus.
Jeff and Eric are 'eagerly' awaiting your comments, feedback, and (of course) questions. I think RMark is a superb addition to the analysis of data from marked individuals, and woul like to extend my congratulations to Jeff and Eric for their efforts.
http://www.phidot.org/software/mark/download/RMark.zip
or from Jeff's NOAA site.
In addition, Jeff and Eric have recently completed a comprehensive 60+ page 'primer' on using RMark with MARK - it cmprises the new 'Appendix C' for 'the book' - and can be downloaded from the usual sources, or directly here at
http://www.phidot.org/software/mark/docs/book/app_3.pdf
Why should you look at RMark? Well, as most experienced MARK users will confirm, there will come a time when your models are so big, and so complex, that the thought of building a design matrix manually fills you with something approaching dread. You might even end up saying a 'bad word' (or several) if your DM doesn't work. So, in part to simplify your life, RMark was designed to let you build complex models without ever having to 'play' with the design matrix. Using a natural langauge approach to building models within the RMark library, you can quickly build even complex models for most of the available data types currently available in MARK, using a fairly straightforward model building syntax. The use of a natural language approach to building models is not new - it exists to some degree in POPAN - and very much so in M-SURGE. However, the RMark approach also gives you access to all the other functionality of R, which is a definite plus.
Jeff and Eric are 'eagerly' awaiting your comments, feedback, and (of course) questions. I think RMark is a superb addition to the analysis of data from marked individuals, and woul like to extend my congratulations to Jeff and Eric for their efforts.