Machine Translation I — EKKTT:
The project is a continuation of the project "Machine Translation I" that started in the year 2004 as a part of the National Programme for Estonian Language and National Memory. As a part of the project, a sentence-aligned English-Estonian parallel corpus has been created. In addition, first experiments of Estonian-English statistical machine translation have been conducted.
'via Blog this'
Estonian-English Statistical Machine Translation: the First Results
ReplyDeleteAbstract
This paper describes the experiments that
apply phrase-based statistical machine
translation to Estonian. The work has two
main aims: the first one is to define the
main problems in the output of EstonianEnglish statistical machine translation and
set a baseline for further experiments with
this language pair. The second is to compare the two available corpora of translated
legislation texts and test them for compatibility. The experiment results show that
statistical machine translation works well
with that kind of text. The corpora appear
to be compatible, and their combining –
beneficial
http://dspace.utlib.ee/dspace/bitstream/handle/10062/2589/post-Fishel-13.pdf;jsessionid=58E4F5A1E1D06E692F07D1DD79EEF1DC?sequence=1