Whenever I lemmatize a compound word in English or German, I obtain a result that ignores the compound structure, e.g. for 'sidekicks' the NLTK WordNet lemmatizer returns 'sidekick', for 'Eisenbahnfahrer' the result of the NLTK German Snowball lemmatizer is 'eisenbahnfahr'. What I need, however, is something that would extract the primary components out of compound words: ['side', 'kick'] and, especially, ['eisen', 'bahn', 'fahr'] (or 'fahren' or in whatever form for the last item). I am especially interested in segmentizing compound words for German.

I failed to find anything of the kind. This kind of an NLP pipe would probably not be called a lemmatizer (or would it?) Is there a definition for it?



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.