R help - Jul 2003 - how to handle missing values

This group impresses me, so far I have been helped with all my questions
within 24 hours. Thanks.

Therefore another one.

I am used to programs (such as STATA) where observations with missing values
that are included in a model are simply ignored in the analysis. So far I
have not been able to figure out how to deal with missing values in R and
have solved the problem by deleting observations with missing values before
loading them into R.

Can anyone give me a hint on how to do this in a simpler way?

Sincerely,


Dr. Tor A Strand   
Centre for International Health
Haukeland Hospital
University of Bergen
5021 Bergen    
Norway     
Phone: (country prefix 47)
Residence:56 51 10 88, office: 55 97 49 80,
fax: 55 97 49 79, cellular:  90 97 10 86

On Wednesday 16 July 2003 12:11, Tor A  Strand wrote:
> This group impresses me, so far I have been helped with all my
> questions within 24 hours. Thanks.
>
> Therefore another one.
>
> I am used to programs (such as STATA) where observations with
> missing values that are included in a model are simply ignored in
> the analysis. So far I have not been able to figure out how to deal
> with missing values in R and have solved the problem by deleting
> observations with missing values before loading them into R.
>
> Can anyone give me a hint on how to do this in a simpler way?
R can deal with NAs in several ways, look at
  help(na.omit)

Many methods, e.g., lm(), also take an na.action argument so that you 
can keep your data frame with the NAs in R, but fit linear models only 
on the complete cases for example.

Best,
Z

> Sincerely,
>
>
> Dr. Tor A Strand
> Centre for International Health
> Haukeland Hospital
> University of Bergen
> 5021 Bergen
> Norway
> Phone: (country prefix 47)
> Residence:56 51 10 88, office: 55 97 49 80,
> fax: 55 97 49 79, cellular:  90 97 10 86
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help

There are many packages that handles missing values, aka imputations.
>From memory Hmisc is one of them.  But you can certainly search through 
CRAN.

On Wed, 16 Jul 2003, Tor A  Strand wrote:
> Date: Wed, 16 Jul 2003 12:11:32 +0200
> From: Tor A  Strand <Tor.Strand at cih.uib.no>
> To: R-list <r-help at stat.math.ethz.ch>
> Subject: [R] how to handle missing values
> 
> This group impresses me, so far I have been helped with all my questions
> within 24 hours. Thanks.
> 
> Therefore another one.
> 
> I am used to programs (such as STATA) where observations with missing
values
> that are included in a model are simply ignored in the analysis. So far I
> have not been able to figure out how to deal with missing values in R and
> have solved the problem by deleting observations with missing values before
> loading them into R.
> 
> Can anyone give me a hint on how to do this in a simpler way?
> 
> 
-- 
Cheers,

Kevin

------------------------------------------------------------------------------
"On two occasions, I have been asked [by members of Parliament],
'Pray, Mr. Babbage, if you put into the machine wrong figures, will
the right answers come out?' I am not able to rightly apprehend the
kind of confusion of ideas that could provoke such a question."

-- Charles Babbage (1791-1871) 
---- From Computer Stupidities: http://rinkworks.com/stupid/

--
Ko-Kang Kevin Wang
Master of Science (MSc) Student
SLC Tutor and Lab Demonstrator
Department of Statistics
University of Auckland
New Zealand
Homepage: http://www.stat.auckland.ac.nz/~kwan022
Ph: 373-7599
    x88475 (City)
    x88480 (Tamaki)

This really depends on what you want to do. I will try to give some
example below.


1. Coding the missing values

But you definitely do not need to delete observations BEFORE loading
them into R. 
By default any empty cells or "NA" is treated as NA, when you load the
data using read.delim(). You can adjust the na.string option in
read.delim() to change this default behaviour.

Ensure the coding is ok before you proceed. You can check using is.na()
for example to see if R will treat them as missing values.


2. Perform calculations with missing values as defined by na.action()

X <- c(1,2,3,NA, 4)
sum(X, na.rm=T)     # gives you 10
See ?na.action for more interestin detail.

Some algorithms are capable of automatically handle missing values. In
the classification context, rpart can handle missing values.


3. Missing value imputation

There are many imputation methods (eg. EMV, e1071, hmisc, norm, permax,
pamr libraries). The type of imputation depends on your application,
area of research and type of missingness (if at missing completely at
random, missing/observed at random, 
informative missing).

Good luck.


-----Original Message-----
From: Tor A Strand [mailto:Tor.Strand at cih.uib.no] 
Sent: Wednesday, July 16, 2003 6:12 PM
To: R-list
Subject: [R] how to handle missing values


This group impresses me, so far I have been helped with all my questions
within 24 hours. Thanks.

Therefore another one.

I am used to programs (such as STATA) where observations with missing
values that are included in a model are simply ignored in the analysis.
So far I have not been able to figure out how to deal with missing
values in R and have solved the problem by deleting observations with
missing values before loading them into R.

Can anyone give me a hint on how to do this in a simpler way?

Sincerely,


Dr. Tor A Strand   
Centre for International Health
Haukeland Hospital
University of Bergen
5021 Bergen    
Norway     
Phone: (country prefix 47)
Residence:56 51 10 88, office: 55 97 49 80,
fax: 55 97 49 79, cellular:  90 97 10 86

______________________________________________
R-help at stat.math.ethz.ch mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help

Thank you for all the responses on generalized additive models(gam) and
missing values. I am now able set up a model using gam and have a certain
understanding of how R deals with missing values.

The problem is, however, I am still not able to a gam model that is from a
dataset that contains missing values.

The function

C<-gam(depvar~var1+var2+s(var3), data=dataset)
Returns the errors

Error in na.omit.default() : Argument "object" is missing, with no
default

Again, can anyone help a newbie.

Tor