[Traminer-users] How to treat in different ways different types of missing values?

Elie Chosson elie.chosson at upmf-grenoble.fr
Tue Sep 25 12:13:46 CEST 2012


Hi,

thank you very much!
Decidedly, Traminer has all the answers! I'm going to see that now...

Sincerly,
Elie

Le 25/09/2012 12:00, traminer-users-request at lists.r-forge.r-project.org 
a écrit :
> Hi Elie,
>
> There is a way of handling missing values differently depending on where
> they appear in the sequences by using the "left", "right" and "gaps"
> options when defining your sequence object with 'seqdef'. And it seems
> that you can solve your problem by setting these options.
>
> Everything is explained in details in the manual page of 'seqdef', in
> the user's guide and in our paper published in the Journal of
> Statistical Software. All this documentation that we wrote especially
> for you the TraMineR users is freely available on our website or
> directly from within R.
>
> All the best,
> Alexis
>
> Le 21. 09. 12 11:04, Elie Chosson a ?crit :
>> >  Hello, dear TraMineR's users!
>> >
>> >  I'm preparing my Ph.D in economy, and I use for it R and the TraMineR
>> >  package. It's a powerful, useful and very nice tool!
>> >  I work on the trajectories of the recipients of the french "Revenu de
>> >  Solidarit? Active". I have almost 2000 trajectories in my data frame,
>> >  and my data cover 9 quarter.
>> >
>> >  My problem is the next:
>> >  I have some missing values, but with a sens which is not all times the
>> >  same:
>> >  1. first case: when a missing value for a quarter is between two
>> >  normal values for the previous and next quarter.
>> >  2. second case: a missing value for a quarter is followed by missing
>> >  values for all the next quarters.
>> >
>> >  example:
>> >  _first case:_
>> >  Normal-Normal-Normal-*NA*-Normal-Normal-Normal-Normal-Normal
>> >  _second case:_  Normal-Normal-*NA*-*NA*-*NA-NA*-*NA*-*NA*-*NA*
>> >
>> >  In the first case I would like to treat the NA state in the same way
>> >  than the following state, and in the second case the NA state is
>> >  correct, and I can consider it like a outing of the policy than i study.
>> >
>> >  In your mind, their is possibilities with TraMineR to do this? In the
>> >  seqdef function, can I specify a rule like it? May be in defining the
>> >  labels and the codes? Can I change a state in a trajectory with a
>> >  logical function?
>> >
>> >  I don't know if I'm clear, and i sorry for my bad English,
>> >
>> >  thanking you in advance,
>> >  -- 
>> >  Elie Chosson,
>> >  Doctorant contractuel,
>> >  Centre de Recherche en ?conomie de Grenoble (CREG)
>> >  Bureau 513
>> >  BP 47 - 38040 Grenoble Cedex 9
>> >
>> >  tel: 04 76 82 59 89 / 06 43 70 99 29
>> >
>> >
>> >  _______________________________________________
>> >  Traminer-users mailing list
>> >  Traminer-users at lists.r-forge.r-project.org
>> >  https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:<http://lists.r-forge.r-project.org/pipermail/traminer-users/attachments/20120924/1f8a7cdc/attachment-0001.html>
>
> ------------------------------
>
> _______________________________________________
> Traminer-users mailing list
> Traminer-users at lists.r-forge.r-project.org
> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users
>
> End of Traminer-users Digest, Vol 25, Issue 8
> *********************************************
>


-- 
Elie Chosson,
Doctorant contractuel,
Centre de Recherche en Économie de Grenoble (CREG)
Bureau 529
BP 47 - 38040 Grenoble Cedex 9

tel: 04 76 82 54 50

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/traminer-users/attachments/20120925/4a81b523/attachment.html>


More information about the Traminer-users mailing list