| about | publications | software | links |
QTag is a freely available, language independent POS-Tagger. It is implemented in Java, and has been successfully tested on Mac OS X, Linux, and Windows. While it works in principle with any language it has resource files for, the ‘shrink-wrap’ release comes only with data for English.
While creating resource files for other languages (from a pre-tagged training corpus) is a fairly trivial task, it has not yet been integrated into a user-friendly procedure, mainly because it is something that happens not frequently enough. If you have a tagged corpus that you would like to use for training QTag, please get in touch with me.
With my current set of webpages being re-worked following changes at the University, the documentation is also not yet fully finished. If you have any questions, just ask!
QTag comes as a Java archive file. It has a basic GUI included, so you can simply run it with
java -jar qtag.jar
(you need the Java Runtime Environment, at least version 1.2). Everything else should be self-explanatory.
Other people have used QTag for their research; I am always pleased to see that my software is useful. But I was especially pleased to read the following passage about a comparison of three POS-taggers (thanks to Dominic Hughes for bringing this to my attention):
QTag was found to be the fastest (by at least an order of magnitude compared to the slowest) and most robust when dealing with misspelt words and other ‘junk’ text. There was little difference in accuracy between the three.
This is from Johnson, D; Malhotra, V; Vamplew, P More Effective Web Search Using Bigrams and Trigrams, Webology, Volume 3, Number 4, December, 2006.
Further software will be added as it is ready. Currently planned:
If you have used my software and you like it, you can contribute to me developing other software you might also like. Donating some money (as much as you like) will help me to buy more programming books, for example. You will in turn get my gratitude and some useful software as well!