[R] Package norm has been removed. What to use for Maximum Likelihood Missing Data Imputation?

KathyKlein kleinkatharina at gmx.net
Wed Jul 29 15:08:35 CEST 2009


Hello,

I also used the norm-package for data with missing values. Are there no
possibilities to fix it (its not under GPL?), or is here anyone able to get
the negotiation for the maintainership for this package?

Another question: I have read, taht "mice" would be an alternative instead
of using "norm". Is it also possible to the EM-Algorithm for imputation of
missing values in multivariate normal data? Or are there better
possibilities?
(Besides:the package "mvnmle" is not possible for me, because my data matrix
is too big...)

I would be pleased to get anwers from you!

Katharina Klein



Uwe Ligges-3 wrote:
> 
> 
> Achim Zeileis wrote:
>> On Fri, 17 Jul 2009, Daniel Abbott wrote:
>> 
>>> Hello,
>>>
>>> I apologize if an answer to my questions is available, or if I
>>> submitted this question incorrectly. I have read the mailing lists, as
>>> well as the R Project and CRAN homepages. However, I may have missed
>>> something.
>>>
>>> I noticed the package 'norm' has been removed. Its page
>>> http://cran.r-project.org/web/packages/norm/index.html now reads:
>>>
>>> "Package ?norm? was removed from the CRAN repository.
>>>
>>> Formerly available versions can be obtained from the archive."
>>>
>>> My questions are:
>>>
>>> 1. Why was norm removed? I used the package and found it to be very
>>> useful. Is there a serious error in the package or another problem?
>> 
>> What do you consider to be a serious error?
>> 
>> Typical reasons for removing a package from the active CRAN repository 
>> are that the package does not pass CRAN checks and/or that the package 
>> maintainer is irresponsive. (norm, specifically, hadn't been updated 
>> since 2002.) This is serious enough to make all automatic CRAN features 
>> such as daily package checks and building of binary packages too 
>> cumbersome.
>> 
>> But if you do not consider this to be serious, you can keep on using 
>> "norm", it is still in CRAN's package archives. Or, even better, you 
>> could adopt it, fix it, and release a new version to CRAN (although it 
>> is not quite clear to me whether the norm's license allows this).
> 
> 
> Oh yes, indeed.
> 
> Please ignore my former message, I haven't looked at the license 
> carefully enough. And since the license is rather special (not GPL'ed as 
> I assumed in my message), you cannot take over maintainership without 
> negotiation with the former maintainer, I fear.
> 
> Best,
> Uwe Ligges
> 
> 
> 
>> (Note that CRAN runs no checks whether the methods implemented are 
>> reasonable or well-suited for the problem they are trying to address
>> etc.)
>> 
>>> 2. If norm should no longer be removed, what is another good package
>>> for ML (maximum likelihood) missing data imputation? I have seen
>>> several recommendations online, but I wonder which one is currently
>>> the "reigning champ."
>> 
>> The "Multivariate" and the "SocialSciences" task views have sections 
>> about missing data
>>   http://CRAN.R-project.org/view=Multivariate
>>   http://CRAN.R-project.org/view=SocialSciences
>> 
>> Additionally, I can recall that there is the Amelia II package:
>>   http://CRAN.R-project.org/package=Amelia
>> and potentially others.
>> 
>> hth,
>> Z
>> 
>>> Thank you very much. I appreciate your time!
>>>
>>> ______________________________________________
>>> R-help at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-help
>>> PLEASE do read the posting guide 
>>> http://www.R-project.org/posting-guide.html
>>> and provide commented, minimal, self-contained, reproducible code.
>>>
>>>
>> 
>> ______________________________________________
>> R-help at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide 
>> http://www.R-project.org/posting-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
> 
> 

-- 
View this message in context: http://www.nabble.com/Package-norm-has-been-removed.-What-to-use-for-Maximum-Likelihood-Missing-Data-Imputation--tp24531976p24718866.html
Sent from the R help mailing list archive at Nabble.com.




More information about the R-help mailing list