[S] nls() finds mystery NA's?

[Joe Mortzheim <jmortz@snake1.cr.usgs.gov>]

    I am calling nls using the following code:

-----------------------------------------
   Output <- nls(
    The.Original.Formula,
    Temp.Regression.Data,
    start=list(B1=Temp.B1,B2=Temp.B2,B3=Temp.B3,B4=Temp.B4,B5=Temp.B5),
    trace=T)
-----------------------------------------

        Very often the nls() routine will fail to completely execute and
dump.  It gives me a message indicating that the nls() routine found one
or more NA observations in the dataset.  Observing the trace output I
see that the nls() routine iterates from one to many times.

    Sometimes the error indicates that the nls() routine found a single
NA.  For example:

-----------------------------------------
Error in nls(The.Original.Formula, Temp.Regression.Data, ..: NA found in
value
-----------------------------------------

    Other times the error indicates that the nls() routine found a many
NA observations.  In these instances it is even more frustrating because
it tells me that ALL of my observations were NA.  For example, I might
send 101 observations and receive the following error message:

-----------------------------------------
Error in nls(The.Original.Formula, Temp.Regression.Data, ..: 101 NAs
found in value
-----------------------------------------

    I am certain that the data going in to the nls() routine (a.k.a.
Temp.Regression.Data) contains no NA observations and most definitely it
cannot have 101 NA observations.  What is going on?  This problem has me
stumped.  I am using S+ 3.3 under WindowsNT.

    Thanks in advance for your help.

    Joe


-----------------------------------------------------------------------
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