Notebooks

Computational Statistics

20 Dec 2011 21:07

By this I do not just mean R, but R is a big part of being a working academic statistician these days...

R, for the record, is a free, open-source interpreted programming language (and interactive environment) for statistical computing. It descends from a language developed at Bell Labs (of blessed memory) called S. There is a commercial descendant of S called S-plus, but I know of no reason to use it, rather than R. For that matter, I know of no reason to use any of the commercial statistical environments (Stata, SPSS, Minitab, ...) rather than R, except for pesonal and organizational inertia. (Which is not to be slighted, of course.) The only real alternative, from my point of view, is hand-written code in something like C/C++ or Fortran --- which can of course be integrated with R. It would be a bit unfair to say that seeing a new method without an R implementation is cause for suspicion, but not wildly unfair.

(And, of course, people who use Excel to do statistics are not to be taken seriously.)

See also: Statistics; Monte Carlo; Data Mining; Programming


Notebooks:     Hosted, but not endorsed, by the Center for the Study of Complex Systems