Mining comparable bilingual text corpora for cross-language information integration