Hello, I am trying to implement a coreference resolving system for Dutch. So far, I already trained a parser model to obtain the noun phrases and I am now in the point of understanding the coreference resolving system.
I have doubts about how the Wordnet dictionary is used. My understanding is that it is used only to obtain synonyms: - For each noun phrase, the head of the phrase is obtained. - For the head, Wordnet is used to get the list of synonym senses: -- First, the lemma of the word is obtained. -- Second, the synonym sense for the word in a noun context is obtained - i.e, a number representing a meaning of the word when it is used as a noun. -- Third the parents' synonym senses are obtained. What are exactly these parents? Viewing the code I think they are the rest of the words of the same type (noun) that are considered hypernyms - i.e, if the word is seagull, we will get the the synonym sense for "bird" also, and recursively the synonym sense "animal" Is this right and complete, or am I missing or misunderstanding some use of the Wordnet dictionary? Is it also used at coreference training time? Best regards! Ismael. -- Ismael Hasan Romero http://www.linkedin.com/in/ismaelhasan/en
