ABSTRACT - DELIVERABLE 106
Report on analogue matching software and package 'analogue'p> This deliverable is prepared in three main parts. The first provides a short in− troduction to the analogue package for the R statistical software. The second part is a draft manuscript submitted to the Journal of Statistical Software de− tailing the usage of the analogue package via a worked example. The final part of this deliverable is the documentation for the package functions themselves.
It was intended that this deliverable would report on an application of analogue matching of a Scottish data set of MAGIC hydrochemistry hindcasts and fore− casts to a modern training set of hydrochemistry data available for the UK. This work has been completed and will be presented at the Leipzig project meeting April 2007. However, during the course of performing this analysis and other work under WP8, it was realised that a stable implementation of the code required to perform the analytical methods was required if other project participants were to be able to undertake analogue matching on their data sets. Work on a package of R code that implemented analogue matching and related methods was, as a result, begun. An opportunity to produce a manuscript for a leading journal dealing with statistical software arose, and in order to meet the journal deadline, resources were diverted away from writing the report on the Scottish analogue matching study and into finalising the analogue package and writing the accompanying manuscript. As such we have delayed the reporting of the Scottish study until the Leipzig project meeting and a subsequent deliverable. In its place, we provide a copy of the draft manuscript detailing the use of the analogue package as well as the software manual.
R is a language and environment for statistical computing and graphics. It is a GNU project which is similar to the S language and environment which was developed at Bell Laboratories (formerly AT&T, now Lucent Technologies) by John Chambers and colleagues. R can be considered as a different implemen− tation of S. There are some important differences, but much code written for S runs unaltered under R.
The main reason for choosing to write the analogue package for R is that R is open source, released under the GNU General Public Licence (GNU GPL), and available freely to anyone who wants to use it. R also provides a rich source of programming tools that allow high quality output and graphical displays to be produced at relatively little effort.
Like R, analogue is also made available under the GNU GPL and as such is freely available and the source code can be consulted to understand how the package works and to modify it if the user so wishes to added new functionality or modify the implementation.analogue has been submitted to the Comprehensive R Archive Network (CRAN), an on line repository for R packages. By making analogue available via CRAN, anyone who has R installed on their computer can install analogue with minimal effort using the package management tools built into R. analogue implements novel and cutting edge topics in the use of modern analogues for environmental modelling, that are not available in any other com− puter package or application. By drawing together current research tools into a software package for end users we gain the input of many users in how these functions perform. This input will also feedback into the development of the techniques that is ongoing as part of research within Euro−limpacs and beyond.
This report contains documentation on how to use package analogue including a draft manuscript submitted to the Journal of Statistical Software that contains a detailed worked example.
Please login to download this report