SQ-Ados for Sequence Analysis updated on SSC

[Ulrich Kohler]

Dear Statalisters,

Courtesy of Kit Baum, the SQ-Ados for Sequence Analysis has been updated on SSC. Users of previous versions of the SQ-Ados should type

adoupdate sq, update

to update their installation. Stata users who want to install the package the first time should type

ssc install sq

instead.

The following is a short summary new features. The order of the mentions reflect my sense of importance. The most important changes are mentioned first.

1. sqom now has an interface to SADI, Brendan Halpins Stata Plugins for Sequence Analysis. This gives the users access to all distance measures provided by SADI and the advantage of the higher speed of these plugins. In order to invoke SADI, users of the SQ-Ados invoke sqom with the option sadi(name), where name is the name of one of SADI's distance measure. For example, the command
   
   

   	. sqom, sadi(oma)

   creates a distance matrix with the Levenshtein matrix. The distance matrix left behind by sqom with option sadi is indistinguishable from the one that is left behind by standard sqom. This means that it can be directly accessed by the the SQ-Ados interface commands such as sqclusterdat and sqmdsadd.
   
   Users interested in using this function should also install SADI using
   
   

   	. ssc install sadi

2. So far it was recommended not to use wardslinkage for a cluster analysis on a distance matrix created by sqom. This recommendation is no longer valid. I do not see any obstacles against using wardslinkage in connection with the distance matrix left behind by sqom.
3. The SQ-Ados now include programs to deal with sequences stored in string variables. Particularly, the new egen function sqstrnn(varname) searches in varname for strings that are similar to each entry in varname and writes the result to a new variable. This can be helpful for harmonizing the spellings of words in a string variable. Secondly the command sqstrlev varname creates a matrix holding the pairwise Levenshtein distances of all strings in varname. And, last but not least, the command sqstrmerge can be used to merge two datasets using a key variable with entries that are similar, but not identical.
4. sqpercentageplot is a new graphical display of the sequence data set
5. The egen function sqsucess() measures the sucess of a binary sequence

[Julia Peters]

Dear Ulrich,

thank you very much for this updated version.

I am trying to run an Optimal Matching Analysis with this new update on a Mac, with Stata 14.1.
I have installed this updated version and also tried updating it (just in case). Everything seems fine and up to date.

Also, all descriptive commands work fine (sqtab, sqdes, sqindexplot etc.).

But whenever I am trying to run the sqom, I receive the following error:

sqomref() not found

I have read older posts with the same issue (e.g. http://www.stata.com/statalist/archi.../msg01316.html), but cannot seem to figure out what the problem is.

Is there anything else I need to download?

Or is this a Mac-specific problem?

Thank you very much!

Best,
Julia

[Ulrich Kohler]

Dear Julia,

Mac shouldn't be a problem here. Unfortunately, however, I do not know what is going on. Perhaps you can "complie" the Mata code yourself. Do do that, please -cd- do the directory in which the file lsq.mata is located and run this as file as do-file:

. cd `c(sysdir_plus)'/l
. do lsq,mata

Afterwards -cd- back to your working directory and try again. Drop me a note if you still receive any errors.

Many regards
Uli

[Anna Manzoni]

I am getting this same error with sqom (sqomref() not found) after I updated and I am unable to use sqom anymore.... I could not follow the suggestion about the Mata code in the previous reply. Any further suggestion about how to solve this is greatly appreciated. Thank you

[Anna Manzoni]

I managed to recompile the Mata library and it now works fine.

[jatta salmela]

Dear Ulrich,

I am doing sequence analysis with Stata 15.1., and I would like to use sqpercentageplot command. I have installed an updated version (adoupdate sq, update), and everything is working fine (e.g., sqindexplot, sqom). Also, I managed to install sadi. However, when running the command sqpercentageplot, it is all the time "unrecognized".

Do you have any suggestions/ideas, why I cannot run sqpercentageplot? Is it because of not having the latest Stata version (16)?

Kind regards,
Jatta Salmela