st: SQ Package updated on SSC

[Ulrich Kohler <kohler@wzb.eu>]

<>

I have sent to Kit Baum a new program, -sqmodalplot-, which belongs to
the SQ-package for Sequence analysis. The program will be downloadable
from SSC in due time. Users of the SQ-package are asked to update their
installation with 

. adoupdate sq, update

 Users who want to try the SQ-package the first time may install the
package with

. ssc install sq

The new program draws a plot of the modal sequence over specified
groups. A modal sequence is an artificial sequence composed by the most
frequent element at each position. Think of the modal sequence as some
form of an ideal-typical sequence, which implies that the modal sequence
do not necessarily exist as a whole in the data set.

-sqmodalplot- is useful to show the results of a cluster analysis on the
distance matrix created by -sqom, full-. Another use is to inform the
user about possible settings for the option -idealtype()- of -sqom-.
Finally -sqmodalplot- provides a way to prevent the tiresome
overplotting-problem of -sqindexplot-. I like the idea behind
-sqmodalplot- very much.

I take the opportunity to inform the users of the SQ package about some
of the more recent updates for the SQ package. All these updates have
been distributed solely via the SSC-Archive and _not_ as an update of
the original Stata Journal distribution. I therefore recommend to
install the SQ-package from SSC. The Stata Journal package -st0111- is
out of date. 


-sqmdsadd- to add results of MDS to sequence data
-------------------------------------------------

The new program -sqmdsadd- adds the result of a Multi Dimensional
Scaling on the dissimilarity Matrix created by -sqom- to the sequence
data. In my experience, Multidimensional Scaling is much more powerful
for sorting the sequences for a sequence index plot than the more
commonly used cluster analysis. 

-sqom- for Optimal Matching
---------------------------

The option -idealtype()- of -sqom- can be used to specify an artificial
reference sequence with which all the sequences in the data are
compared.

In addition, -sqom- now allows two subcommands, -sqom use- and -sqom
save-, to store and re-use the distance matrices created by -sqom- 


New functions for -sqegen-
--------------------------

Several new functions for egen have been created (see -help sqegen-): 

-sqfirstpos()- creates a variable holding the position where a specified
subsequence is first found. 

-sqallpos()- creates a variable hold the number of positions at which a
specified subsequence is found. 

-sqfreq()- creates a variable showing how often each type of sequences
appear in the dataset (i.e. the result of -sqtab- as a variable).

-sqranks()- creates a variable showing the rank of a sequence in the
league table of the most frequent sequences (i.e. the position of a
sequence in -sqtab- as a variable). 


-sqindexplot- for Sequence Index Plots
--------------------------------------

Option -order()- allows varlist. 

I wish to thank the users of the SQ package for their requests and bug
reports. 


Finally allow me to announce that the February issue of Sociological
Methods and Research is a special issue on Sequence Analysis. The issue
has been edited by Christian Brzinsky-Fay and myself and features papers
by Brandan Halpin, Laurent Lesnard, Cees Elzinga, Silke Aisenbrey and
Anette Fasang, and Michael Anyadike-Danes and Duncan McVicar. 

http://smr.sagepub.com/content/38/3.toc


Many regards

Uli


-- 
Ulrich Kohler
<kohler@wzb.eu>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/