[S] problems with nls

niels waller (ngwaller@ucdavis.edu)
Sat, 18 Apr 1998 09:45:36 +0100


J Philip Miller has kindly informed me that my previous post contained HTML
code. I apologize if it was unreadable. A cleaned up version of my request
follows:

Problem with nls:

Hello S-Plus enthusiasts.

I am running a very large simulation study in which I call the nls
function. On most data sets (simulated) everything works fine.
However, every now and then nls returns the following error message:
NA values encountered", then it bombs and I lose all of my results (from
the previous 100+ runs). My frustration stems from the fact that I know
that the data do not contain NAs because (1) I generate the data and (2)
I have looked at the offending data (that crashes the program) and it
looks ok. Does anyone have any suggestions. (I supply good start
values and use all other program defaults). I am panicked because I
need to run about 5,000 runs in the next week to prepare for a talk at
NYU.

Thank you for any and all help.

Niels Waller
ngwaller@ucdavis.edu

-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news