<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Hi Alexandre,<br>
<br>
No you should not have a % at the end since
"contact[count.miss>0, 2:4]" displays the input data, not the
sequence object.<br>
<br>
To display the void elements in your sequence object, type<br>
<br>
print(contact.seq, extended=TRUE)<br>
<br>
By "the way your sequences are prepared" I mean that if one or
several missing states (supposed to be NA's by default, but this can
be overridden with "missing=" argument) are found in your input
data, seqdef enters a data preparation stage. The way your missing
values are handled in this stage is set with "left", "gaps" and
"right" arguments. <br>
<br>
Have a look at the way seqdef is handling missing values by using
the ex1 example data set:<br>
<br>
data(ex1)<br>
ex1<br>
seqdef(ex1, 1:13)<br>
seqdef(ex1, 1:13, right=NA)<br>
<br>
All the best,<br>
Alexis.<br>
<br>
Le 20. 04. 11 18:01, Alexandre Pollien a écrit :
<blockquote cite="mid:4DAF0351.8000307@unil.ch" type="cite">Thank
you for your answers Matthias and Alexis,
<br>
<br>
Sorry but I still have a problem with the missings: I can not
understand why my missings are not coded as void element. They are
indeed located after the last valid state and I even tried to
specify the default value right="DEL". I'm especially surprised
because I made this observation on old scripts that did not
react like this before (I think... maybe I'm going mad?)
<br>
<br>
Alexis, what do you mean by "the way your sequences are prepared"?
I usually use csv file
<br>
<br>
An exemple to be clearer:
<br>
<br>
id,c1,c2,c3
<br>
21,8,1,
<br>
<br>
--> and import in TraMineR:
<br>
<br>
<br>
#########################
<br>
<br>
library(TraMineR)
<br>
library(foreign)
<br>
<br>
contact<-read.csv("base.csv")
<br>
<br>
names(contact)
<br>
[1] "id" "c1" "c2" "c3"
<br>
<br>
contact.seq <- seqdef(contact, 2:4, right="DEL")
<br>
[>] found missing values ('NA') in sequence data
<br>
[>] preparing 1 sequences
<br>
[>] coding void elements with '%' and missing values with '*'
<br>
[>] 2 distinct states appear in the data:
<br>
1 = 1
<br>
2 = 8
<br>
[>] alphabet (state labels):
<br>
1 = 1 (1)
<br>
2 = 8 (8)
<br>
[>] 1 sequences in the data set
<br>
[>] min/max sequence length: 2/2
<br>
<br>
count.miss <- rowSums(is.na(contact[,2:4]))
<br>
<br>
contact[count.miss>0, 2:4]
<br>
c1 c2 c3
<br>
1 8 1 NA
<br>
<br>
#########################
<br>
<br>
Should I not have a % at the end ?
<br>
<br>
<br>
<br>
Best (and beautifull easter)
<br>
<br>
<br>
Alexandre
<br>
<br>
<br>
<br>
<br>
<br>
<br>
Le 19.04.2011 12:00, <a class="moz-txt-link-abbreviated" href="mailto:traminer-users-request@r-forge.wu-wien.ac.at">traminer-users-request@r-forge.wu-wien.ac.at</a>
a écrit :
<br>
<blockquote type="cite">Send Traminer-users mailing list
submissions to
<br>
<a class="moz-txt-link-abbreviated" href="mailto:traminer-users@lists.r-forge.r-project.org">traminer-users@lists.r-forge.r-project.org</a>
<br>
<br>
To subscribe or unsubscribe via the World Wide Web, visit
<br>
<a class="moz-txt-link-freetext" href="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users</a>
<br>
<br>
or, via email, send a message with subject or body 'help' to
<br>
<a class="moz-txt-link-abbreviated" href="mailto:traminer-users-request@lists.r-forge.r-project.org">traminer-users-request@lists.r-forge.r-project.org</a>
<br>
<br>
You can reach the person managing the list at
<br>
<a class="moz-txt-link-abbreviated" href="mailto:traminer-users-owner@lists.r-forge.r-project.org">traminer-users-owner@lists.r-forge.r-project.org</a>
<br>
<br>
When replying, please edit your Subject line so it is more
specific
<br>
than "Re: Contents of Traminer-users digest..."
<br>
<br>
<br>
Today's Topics:
<br>
<br>
1. Re: The substitution cost matrix is not symmetric
<br>
(Alexis Gabadinho)
<br>
<br>
<br>
----------------------------------------------------------------------
<br>
<br>
Message: 1
<br>
Date: Mon, 18 Apr 2011 13:54:12 +0200
<br>
From: Alexis Gabadinho<a class="moz-txt-link-rfc2396E" href="mailto:Alexis.Gabadinho@unige.ch"><Alexis.Gabadinho@unige.ch></a>
<br>
Subject: Re: [Traminer-users] The substitution cost matrix is
not
<br>
symmetric
<br>
To: Users questions<a class="moz-txt-link-rfc2396E" href="mailto:traminer-users@r-forge.wu-wien.ac.at"><traminer-users@r-forge.wu-wien.ac.at></a>
<br>
Message-ID:<a class="moz-txt-link-rfc2396E" href="mailto:4DAC2664.3090502@unige.ch"><4DAC2664.3090502@unige.ch></a>
<br>
Content-Type: text/plain; charset=windows-1252; format=flowed
<br>
<br>
Hi Alexandre,
<br>
<br>
Regarding your second question, you have indeed missing values
in your
<br>
data, since we can see in the output of seqdist that the length
of your
<br>
sequences vary from 1 to 12.
<br>
<br>
The way missing values (supposed to be NA's by default) are
handled is
<br>
set in seqdef with the arguments "gaps", "left" and "right". By
<br>
default, missing values found after the last valid state in a
sequence
<br>
are coded as void elements, meaning that your sequence is
supposed to
<br>
end with the last valid state. Thus if you have sequences of
unequal
<br>
length, ending with NA'S , this is the way your sequences are
prepared.
<br>
See user's guide for more information.
<br>
<br>
To locate missing values in your data you can do for example:
<br>
<br>
count.miss<- rowSums(is.na(contact[,2:15]))
<br>
<br>
This counts the number of missing values in each row of your
original data.
<br>
<br>
To display the rows containing at least one missing value:
<br>
<br>
contact[count.miss>0, 2:15]
<br>
<br>
Best regards,
<br>
Alexis.
<br>
<br>
Alexandre Pollien a ?crit :
<br>
<blockquote type="cite">Hello,
<br>
<br>
When I use seqsubm, now I get this message:
<br>
<br>
<blockquote type="cite"> contact.om<-
<br>
</blockquote>
seqdist(contact.seq,method="OM",indel=2,sm=cout,with.missing=FALSE,
<br>
full.matrix=TRUE)
<br>
[>] 5764 sequences with 9 distinct events/states
<br>
[>] 496 distinct sequences
<br>
[>] min/max sequence length: 1/12
<br>
[>] computing distances using OM metric
<br>
[>] total time: 1.25 secs
<br>
Message d'avis :
<br>
The substitution cost matrix is not symmetric.
<br>
<br>
cout is computed with seqsubm (method="TRATE")
<br>
<br>
<br>
A surprising aspect is that it is not systematic (3 / 4)
<br>
<br>
I thought of a problem of missing value
<br>
<br>
<br>
Another amazing prompt is that seqdef shows this message:
<br>
<blockquote type="cite"> contact.seq<- seqdef(contact,
2:15, states = contact.shortlab,labels
<br>
</blockquote>
= contact.lab)
<br>
[>] found missing values ('NA') in sequence data
<br>
<br>
...even when there is no missing in my data! (How can I seethe
<br>
sequence containing the missing data?)
<br>
<br>
<br>
Does anyone have a idea?
<br>
<br>
Thank you very much
<br>
<br>
Best regards and good weekend
<br>
<br>
Alexandre
<br>
<br>
<br>
<br>
_______________________________________________
<br>
Traminer-users mailing list
<br>
<a class="moz-txt-link-abbreviated" href="mailto:Traminer-users@lists.r-forge.r-project.org">Traminer-users@lists.r-forge.r-project.org</a>
<br>
<a class="moz-txt-link-freetext" href="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users</a>
<br>
<br>
</blockquote>
<br>
<br>
------------------------------
<br>
<br>
_______________________________________________
<br>
Traminer-users mailing list
<br>
<a class="moz-txt-link-abbreviated" href="mailto:Traminer-users@lists.r-forge.r-project.org">Traminer-users@lists.r-forge.r-project.org</a>
<br>
<a class="moz-txt-link-freetext" href="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users</a>
<br>
<br>
<br>
End of Traminer-users Digest, Vol 9, Issue 3
<br>
********************************************
<br>
<br>
</blockquote>
<br>
<br>
<pre wrap="">
<fieldset class="mimeAttachmentHeader"></fieldset>
_______________________________________________
Traminer-users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Traminer-users@lists.r-forge.r-project.org">Traminer-users@lists.r-forge.r-project.org</a>
<a class="moz-txt-link-freetext" href="https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users">https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/traminer-users</a>
</pre>
</blockquote>
<br>
</body>
</html>