[R] aregImpute: beginner's question

[Frank E Harrell Jr]

Anders Schwartz Corr wrote:

. . .
> 
> podb2+propdemocracy+avetrade1984dollars+concentration+cycle+polarity+propmid+terrgainer+
> demgainer+ fedgainer+ popdengainer+ urbpopgainer+ tradeopgainer+
> gdppcgainer+ terrloser+ demloser+ fedloser+ popdenloser+ urbpoploser+
> tradeoploser+ gdppcloser, n.impute=100, defaultLinear=TRUE, data=d)
> Iteration:1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
> 25 26 27 28 29 30 31 32 33 34
> Error in lm.fit.qr.bare(f$tx, f$ty) :
> NA/NaN/Inf in foreign function call (arg 1)

This is probably a singularity.  Remove one variable at a time from the 
formula and re-run aregImpute.  That may help you find a culprit.  Also, 
you may not need 100 imputations.

Frank

> 
>>par(mfrow=c(2,3))
>>plot(f, diagnostics=TRUE, maxn=2)
> 
> 2222222222222222222222> fmi <- fit.mult.impute(y ~
> podb2+propdemocracy+avetrade1984dollars+concentration+cycle+polarity+propmid+terrgainer+
> demgainer+ fedgainer+ popdengainer+ urbpopgainer+ tradeopgainer+
> gdppcgainer+ terrloser+ demloser+ fedloser+ popdenloser+ urbpoploser+
> tradeoploser+ gdppcloser, lm, f,
> +                        data=d)
> Error in impute.transcan(xtrans, imputation = i, data = data, list.out =
> TRUE,  :
>         inconsistant naming of observations led to differing length
> vectors
> 

-- 
Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University