Actually there is a DBpedia for Portuguese, it's just hidden because the mappings configuration may not be as good as the others.
http://downloads.dbpedia.org/3.6/pt/ If you find obvious issues with the existing mappings, don't forget that it's possible to contribute fixes / complements here: http://mappings.dbpedia.org/ The fact that there is some support for Portuguese in DBpedia means that is might be possible to train OpenNLP models for this language as explained in this blog post: http://blogs.nuxeo.com/dev/2011/01/mining-wikipedia-with-hadoop-and-pig-for-natural-language-processing.html Please let me know if you decide to go this way (it's not trivial but might be worth the investment). Also join the OpenNLP mailing list because training NLP corpus is an ongoing issue an every contribution is more than welcomed. -- Olivier
