[Traminer-users] clustering of event sequences

[Matthias Studer]

Dear Hugo,

Thank you for your question. Unfortunately, the function was not 
publicly released and that's why you didn't found it. So please 
apologize for this. I have now added it to TraMineRextras . As you may 
find on the R-forge page 
(http://r-forge.r-project.org/forum/forum.php?forum_id=4296),  "This 
package provides functions that haven't been integrated into TraMineR 
either because they do not completely comply with the TraMineR logic and 
homogeneity requirements or because they need further testing".

You may install this package using the following command. You may need 
to wait 24 hours before the new version of the package is available (you 
may check here https://r-forge.r-project.org/R/?group_id=743 if  Rev. is 
set to 194, it's OK).

install.packages("TraMineRextras", repos="http://R-Forge.R-project.org")

The function to compute distance between event sequence is "seqedist" 
and the function to represent graphically a set of event sequence is 
called seqedplot.

Here is a short example.

  library(TraMineRextras)
library(cluster)
data(actcal.tse)

data(actcal.tse)
actcal.seqe <- seqecreate(actcal.tse)
## We have 8 different event in this dataset
## We set indel cost of an event to 1 and v param to 0.1
idcost <- rep(1, 8)
dd <- seqedist(actcal.seqe, idcost=idcost, vparam=.1)

## These events sequences are mainly stationnary...
cluster10 <- pam(dd, diss=TRUE, k=10, cluster.only=TRUE)
seqedplot(actcal.seqe, group=cluster10, type="hazard", breaks=6)


Hope this helps.
Matthias Studer

Le 16.03.2012 10:45, Hugo Varet a écrit :
> Dear all,
>
> I read with interest the paper from Matthias Studer entitled "Classer, 
> discriminer et visualiser des séquences d'événements" and I discovered 
> I could use exactly the same methods for my study. He said in his 
> paper that all fonctions can be found in the TraMineR package. 
> Unfortunately, I did not find them in the package: I understand how to 
> build homogeneous groups of indivuals from state sequences but not 
> from event sequences (TSE format).
>
> Could anybody indicate me the names of the functions I am looking for ?
>
> Thank you very much, best regards,
>
> Hugo
>
>
> _______________________________________________
> 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/20120316/a87fbc96/attachment.html>