Dear All, Is it possible to perform OLS using C code? I am trying to optimize a n-period "moving window" OLS on a huge dataset hence was wondering if such a thing is possible. Ideally the solution that I am looking for would involve a C-code accepting two float arrays and returning back computed parameters such as t-stat, coefficient etc. I have glanced thru the FAQ's and tried to google this information but wasn't able to find anything truly relevant, hence would really appreciate your help on this one. Thanks. Manoj
Performing regression using R & C
4 messages · Manoj - Hachibushu Capital, Greg Tarpinian, Roger Bivand +1 more
I have found "Numerical Recipes in C" to be extremely
helpful in my statistical studies. OLS is described
in detail in this book.
Best,
Greg
--- Manoj - Hachibushu Capital <Wanzare at HCJP.com>
wrote:
Dear All, Is it possible to perform OLS using C code? I am trying to optimize a n-period "moving window" OLS on a huge dataset hence was wondering if such a thing is possible. Ideally the solution that I am looking for would involve a C-code accepting two float arrays and returning back computed parameters such as t-stat, coefficient etc. I have glanced thru the FAQ's and tried to google this information but wasn't able to find anything truly relevant, hence would really appreciate your help on this one. Thanks. Manoj
______________________________________________ R-help at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
On Thu, 18 Nov 2004, Greg Tarpinian wrote:
I have found "Numerical Recipes in C" to be extremely helpful in my statistical studies. OLS is described in detail in this book.
Maybe just read the R code: src/library/stats/R/lm.R contains function lm.fit(), which is a wrapper for the Fortran subroutine dqrls(), and is about as finely honed as you get. It won't give you the standard errors, but you get the coefficients and the qr object, from which the rest can be constructed. Using well-supported and maintained code does save time too! But maybe there are moving window functions around that you can use without having to "roll your own", searching on "moving window" at http://finzi.psych.upenn.edu/search.html gave 85 hits, including the following, which you may find fruitful: http://finzi.psych.upenn.edu/R/Rhelp02a/archive/26904.html Roger
Best,
Greg
--- Manoj - Hachibushu Capital <Wanzare at HCJP.com>
wrote:
Dear All, Is it possible to perform OLS using C code? I am trying to optimize a n-period "moving window" OLS on a huge dataset hence was wondering if such a thing is possible. Ideally the solution that I am looking for would involve a C-code accepting two float arrays and returning back computed parameters such as t-stat, coefficient etc. I have glanced thru the FAQ's and tried to google this information but wasn't able to find anything truly relevant, hence would really appreciate your help on this one. Thanks. Manoj
______________________________________________ R-help at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
______________________________________________ R-help at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Roger Bivand Economic Geography Section, Department of Economics, Norwegian School of Economics and Business Administration, Breiviksveien 40, N-5045 Bergen, Norway. voice: +47 55 95 93 55; fax +47 55 95 93 93 e-mail: Roger.Bivand at nhh.no
Is it possible to perform OLS using C code? I am trying to optimize a n-period "moving window" OLS on a huge dataset hence was wondering if such a thing is possible. Ideally the solution that I am looking for would involve a C-code accepting two float arrays and returning back computed parameters such as t-stat, coefficient etc.
It is probably not worth trying to call R to do OLS from C. I would look at GSL: http://www.gnu.org/software/gsl/manual/gsl-ref_toc.html GSL has a simple but elegant C interface for doing OLS and weighted least squares. It will give you the covariance matrix of the model parameters, from which you can derive t-stats. Jim M