I was coming here to ask about an eventual possible tweak of MultiPaX, when I noticed about "today"'s new version of MultiPaX, which does not need the Morphological analyzer, so I added a new question: why? What was it needed for, before? Does new version work differently than the previous? Shall I expect different output?
My original question regarded exactly the root of the verbs (which is found by the Morphological Analyzer): I wanted to ask you if there was a means to tweak MultiPaX to annotate also the root of the verb.
Looking forward for your answer, thanks in advance!
Massi
Massi
Morphological Analyzer in MultiPaX pipeline
Some of the parsers come with their own lemmatization/stemming (e.g. RASP-3). The Stanford parser doesn't do this that's why we use the Morphological Analyzer to get the roots of the terms. I changed the MultiPaX code now to check for root annotations first. If they are missing, the strings, as they appear in the text, are used. In general, the root forms are more convenient to work with and compare with each other.
Thanks for your answer! It [5]
Thanks for your answer!
It makes the same job in another way, in fact.
How do you think I could backtrack the root from the MultiPaX output?