R help - Jul 2006 - set the bahavior that R deal with missing values?

Dear Rusers,
The default behavior in R when performing a regression model with missing
values is to exclude any case that contains a
missing value? How could i set the bahavior that R deal with missing values?
e.g.:
exclude cases listwise
exclude cases pairwise
replace with mean

Thanks very much!

-- 
Kind Regards,
Zhi Jie,Zhang ,PHD
Department of Epidemiology
School of Public Health
Fudan University
Tel:86-21-54237149

	[[alternative HTML version deleted]]

Dear Rusers,
 The default behavior in R when performing a regression model with missing
values is to exclude any case that contains a
missing value? How could i set the bahavior that R deal with missing values?
e.g.:
exclude cases listwise
exclude cases pairwise
replace with mean
Thanks very much!

-- 
Kind Regards,
Zhi Jie,Zhang ,PHD
Department of Epidemiology
School of Public Health
Fudan University
Tel:86-21-54237149

	[[alternative HTML version deleted]]

You can deal with the missing values before performing any regression model.
if you dataset is :

 > D
  X1 X2 X3
1 NA  2  4
2  1 NA NA
3  2  2  6
4 NA 34 NA
5  3 NA  7

then the following will replace all NA with the mean value

 > f<-function(col){
+ col[is.na(col)]<-mean(col,na.rm=TRUE)
+ col}
 > sapply(D,f)
     X1       X2       X3
[1,]  2  2.00000 4.000000
[2,]  1 12.66667 5.666667
[3,]  2  2.00000 6.000000
[4,]  2 34.00000 5.666667
[5,]  3 12.66667 7.000000


Hope this is what you want.

zhijie zhang wrote:
>Dear Rusers,
>The default behavior in R when performing a regression model with missing
>values is to exclude any case that contains a
>missing value? How could i set the bahavior that R deal with missing values?
>e.g.:
>exclude cases listwise
>exclude cases pairwise
>replace with mean
>
>Thanks very much!
>
>  
>

-- 
Ferdinand Alimadhi
Programmer / Analyst
Harvard University
The Institute for Quantitative Social Science
(617) 496-0187
falimadhi at iq.harvard.edu
www.iq.harvard.edu

zhijie zhang <epistat <at> gmail.com> writes:
>  The default behavior in R when performing a regression model with missing
> values is to exclude any case that contains a
> missing value? How could i set the bahavior that R deal with missing
values?
Difficult to say, since there is no such thing as a regression model in R, but
certainly more than a few dozens in several packages. Best show a self-contained
runnably sample where you would prefer some na.action (see, for example,
parameter na.action in lm).

Dieter