[R] loess with missing data points

[Gavin Simpson]

On Tue, 2010-10-05 at 13:56 -0700, 1Rnwb wrote:
> Hello List,
> I have a longitudinal samples on multiple individuals, for initial analysis,
> i am doing some plots to see how the variable changes with the age. In some
> of the individuals one or two data points are missing, 
>    IL1Ra              Age.at.Sample.Collection              Subject.ID
>  6.575466           2.004106                                   00709-0
>  NA                    2.162902                                   00709-0
> 5.115421            3.312799                                   00709-0
> 6.119626            3.515400                                   00709-0
> 5.705648            4.156057                                   00709-0
> 7.298887            4.405201                                   00709-0
> 6.195063            4.577686                                   00709-0
> 6.574864            5.541409                                   00709-0
> When I run
> cap<-paste(sub.id,cases2$group[1],sep=" ")
> plot(cases2$IL1Ra~cases2$Age.at.Sample.Collection, ylab="eo_conc",xlab="Age
> at Sample",pch=1, ylim=c(3,10), xlim=c(0,22),main=cap)
> points(cases2$MIP1b~cases2$Age.at.Sample.Collection,pch=2,ylim=c(1.5,10),col="red")
> points(cases2$MCP1~cases2$Age.at.Sample.Collection,,pch=3,
> ylim=c(1.5,10),col="brown")
> lines(predict(loess(cases2$IL1Ra~cases2$Age.at.Sample.Collection))~cases2$Age.at.Sample.Collection,
> lwd=1, lty=1)
> lines(predict(loess(cases2$MIP1b~cases2$Age.at.Sample.Collection))~cases2$Age.at.Sample.Collection,
> col="red", lwd=1, lty=1)
> lines(predict(loess(cases2$MCP1~cases2$Age.at.Sample.Collection))~cases2$Age.at.Sample.Collection,col="brown",
> lwd=1, lty=1)
> 
> I get error
> Error in model.frame.default(formula = predict(loess(cases2$IL1Ra ~
> cases2$Age.at.Sample.Collection)) ~  : 
>   variable lengths differ (found for 'cases2$Age.at.Sample.Collection')
> In addition: There were 50 or more warnings (use warnings() to see the first
> 50)
> 
> I tried using na.action=na.omit, which is not working at all.
> 
> I would appreciate some pointers in this regards to get these charts.
> Thanks

That's the second time this has come up today. Try

na.action = na.exclude

instead of na.omit in you loess() call. na.exclude will remove the NA
from the data used to fit the model, but ensure theye are put back in to
the results extracted from the loess model using the predict() extractor
function.

HTH

G

-- 
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 Dr. Gavin Simpson             [t] +44 (0)20 7679 0522
 ECRC, UCL Geography,          [f] +44 (0)20 7679 0565
 Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
 Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/
 UK. WC1E 6BT.                 [w] http://www.freshwaters.org.uk
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%