GIZA++ is a statical machine translation toolkit that is used to train IBM Models 1-5 and an HMM word alignment model. This package also contains the source for the mkcls tool which generates the word classes necessary for training some of the alignment models.

For more information on the origins of these tools, refer to http://www.fjoch.com/GIZA++.html and http://www.fjoch.com/mkcls.html.

If you make use of GIZA++ for research or commercial purposes, please cite:

• Franz Josef Och, Hermann Ney. "A Systematic Comparison of Various Statistical Alignment Models", Computational Linguistics, volume 29, number 1, pp. 19-51 March 2003.