[R-sig-ME] plot observed vs fitted with model with NAs in data

Ben Bolker bbolker at gmail.com
Wed Mar 5 16:50:24 CET 2014


  I don't see 'na.action' in your code anywhere.

  This works for me:

mymodel = lmer(hp ~ (1| disp) + (1|cyl), data=mydata,
    na.action=na.exclude)
plot(mymodel, hp ~ fitted(.))

  **However**, the real answer is this item from
https://github.com/lme4/lme4/blob/master/inst/NEWS.Rd, under BUG FIXES
for version 1.1-3:

\item \code{fitted} now respects \code{na.action} settings (Github
      issue #149)

We are hoping to release a new version to CRAN very soon, but in the
meantime version 1.1-4 is available at
http://lme4.r-forge.r-project.org/repos , or from github ...

  Ben Bolker


On 14-03-05 10:40 AM, Paul Tanger wrote:
> Thanks for that tip but it didn't seem to help.  I get the same error.  Is
> there a way to tell it not to predict a value for an NA?  Here is a
> reproducible example:
> 
> data(mtcars)
> mydata = mtcars[2:8]
> # insert some NAs
> mydata$hp = apply(mydata[2], 2, function(x) {x[sample(c(1:length(x)),
> floor(length(x)/10)) ] = NA; x} )
> # run model
> mymodel = lmer(hp ~ (1| disp) + (1|cyl), data=mydata)
> # try to plot obs vs fitted
> plot(mymodel, hp ~ fitted(.))
> 
> Note that whatever na.action I attempt, the error remains.
> Maybe there is a simple solution that I'm missing?
> 
> Thanks!
> 
> 
> On Wed, Mar 5, 2014 at 1:24 AM, ONKELINX, Thierry
> <Thierry.ONKELINX at inbo.be>wrote:
> 
>> You probably need to fit your model with the na.action = na.exclude
>> argument. A more detailed answer requires more information. Have a look at
>> http://www.r-project.org/posting-guide.html
>>
>> ir. Thierry Onkelinx
>> Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
>> Forest
>> team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
>> Kliniekstraat 25
>> 1070 Anderlecht
>> Belgium
>> + 32 2 525 02 51
>> + 32 54 43 61 85
>> Thierry.Onkelinx at inbo.be
>> www.inbo.be
>>
>> To call in the statistician after the experiment is done may be no more
>> than asking him to perform a post-mortem examination: he may be able to say
>> what the experiment died of.
>> ~ Sir Ronald Aylmer Fisher
>>
>> The plural of anecdote is not data.
>> ~ Roger Brinner
>>
>> The combination of some data and an aching desire for an answer does not
>> ensure that a reasonable answer can be extracted from a given body of data.
>> ~ John Tukey
>>
>> -----Oorspronkelijk bericht-----
>> Van: r-sig-mixed-models-bounces at r-project.org [mailto:
>> r-sig-mixed-models-bounces at r-project.org] Namens Paul Tanger
>> Verzonden: woensdag 5 maart 2014 8:24
>> Aan: r-sig-mixed-models at r-project.org
>> Onderwerp: [R-sig-ME] plot observed vs fitted with model with NAs in data
>>
>> Hi,
>> I have a model that I fit from data with some NAs in it.. everything is
>> going well until I try to plot observed vs fitted.. and I can figure out
>> how to fix this.. any ideas?  I can see that there are 1049 observed and
>> 1050 fitted.. but can I resolve this without fitting new models after
>> discarding NAs?  I would really rather not do that.
>>
>>> plot(model, resp~ fitted(.))
>> Error in `[<-.data.frame`(`*tmp*`, , ".y", value = c(130.983333333333,  :
>>   replacement has 1050 rows, data has 1049
>>
>>         [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-mixed-models at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>> * * * * * * * * * * * * * D I S C L A I M E R * * * * * * * * * * * * *
>> Dit bericht en eventuele bijlagen geven enkel de visie van de schrijver
>> weer en binden het INBO onder geen enkel beding, zolang dit bericht niet
>> bevestigd is door een geldig ondertekend document.
>> The views expressed in this message and any annex are purely those of the
>> writer and may not be regarded as stating an official position of INBO, as
>> long as the message is not confirmed by a duly signed document.
>>
> 
> 
>



More information about the R-sig-mixed-models mailing list