[R] cannot calculate standard estimate with predict on loess

Saurav Pathak saurav at sas.upenn.edu
Fri May 4 16:52:34 CEST 2012


On 05/04/2012 10:39 AM, David Winsemius wrote:
>
> On May 3, 2012, at 7:10 PM, Saurav Pathak wrote:
>
>> Hi,
>>
>> For some reason I have been unable to use the predict function when I 
>> desire the standard error to be calculated too.  For example, when I 
>> try the following:
>>
>> l<- loess(d~x+y, span=span, se=TRUE)
>> p<- predict(l, se=TRUE)
>>
>
> I don't know what effect the se=TRUE will have in the first call. As 
> far as I can tell there is no such argument to loess(). Could it be 
> that the extranous argument is have an adverse effect on the effort to 
> later use predict (which does have such an argument.)?

Thanks for your reply.  I tried the following with the same error message:

     l <- loess(d~x+y, span=span)
     p <- predict(l, se=TRUE)

Saurav









>>
>> I get the following error message:
>>
>> Error in vector("double", length) : vector size cannot be NA
>> In addition: Warning message:
>> In N * M1 : NAs produced by integer overflow
>>
>>
>> But when I try the following:
>>
>> l<- loess(d~x+y, span=span, se=TRUE)
>> p<- predict(l, se=FALSE)
>>
>>
>> I have no problem, and p$fit gives me the desired fitted values.  
>> Note that the only difference in this piece of code is se=FALSE.
>>
>> My data d is a vector, and x and y are vectors too of the same length.
>>
>> Any help will be greatly appreciated.
>>
>> Thanks,
>> Saurav
>>
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide 
>> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>
> David Winsemius, MD
> Heritage Laboratories
> West Hartford, CT
>



More information about the R-help mailing list