Skip to content

Computational differences in R vs Excel

3 messages · Carl Witthoft, S Ellison, David Winsemius

#
There are several problems here.   The first is that it's rather unlikely you
really need 10-place accuracy to fit your data.  This suggests you may be
doing something inappropriate such as fitting the wrong function or trying
to extrapolate.   Since you haven't explained what "process" you have that
isn't "converging," since clearly the fitting algorithms themselves have
converged.

Next,  you need to understand that both Excel and R have default convergence
tolerances which may not be identical (as well as default iteration limits).

And finally, of course, there's the question of machine precision limits,
although that is less likely to be a culprit in this instance.  


Soham wrote
--
View this message in context: http://r.789695.n4.nabble.com/Computational-differences-in-R-vs-Excel-tp4680726p4680730.html
Sent from the R help mailing list archive at Nabble.com.
#
My machine (win32, R 3.0.1) gives -6.432232266

Something about your machine, or typing, is different.

It shouldn't be R, as the help system implies it uses the same precision and IEEE floating point rules for all platforms....


S Ellison



*******************************************************************
This email and any attachments are confidential. Any use...{{dropped:8}}
#
On Nov 19, 2013, at 6:02 AM, S Ellison wrote:

            
I get the same, at least when printing to the same degree of accuracy:
[1] -6.432232266


Mac OSX 10.7.5, R 3.0.2