| R is
a free software in the GNU-concept. In this case this means
that a community of statisticians maintain the code and continuously
update the programme on a voluntary basis. The programme is
not easy but is extremely flexible
(statisticians can write and adjust to fit their own specific
wishes); adequate for professional statistics and follows the trends
in modern statistics. Matrix calculations are possible. Bugs are
fixed immediately after discovery. The approach is strictly object
oriented (Scheme inspired): everthing is an object: data, matrices,
results, functions etc with "properties" and "methods" and is
classified in "classes" . Programming of own functions and
procedures is possible.
R is similar to the commercially available S-plus. Last programme has a
better user-interface and graphing, but these features come with a
price-tag as in other professional packages. In terms of
statistical capabilities R is often more advanced as statisticians
can write their own analysis methods and distribute it freely to
colleagues. Where to find the software?
- the software can be downloaded R CRAN for windows or
most other platforms (linux, Apple etc.); both source codes and
precompiled binaries (=executables) are
available;
- first read README before downloading (at
May 10, 2005 the
Windows version between the Binaries is SetupR.exe at http://cran.r-project.org/bin/windows/
- mirror
sites are available; in principle a closer site could be
faster
- very efficient (complete version 1.4.1 is +/- 20 Mbytes
depending on installed documentation) and statistically
powerful.
- official documentation is (very) difficult to read...
More documentation (useful for further study and standard
statistical analysis):
Example of libraries
- Ecology package ADE4
as an example
- GIS interface with
GRASS,
ARCinfo and several spatial statistical packages
- Interface to other statistical programmes, SQL and other
databases: see
packages
- Bioinformatics:
microarray analysis etc..
|