Άρση της αμφισημίας με τον Mnemosyne Tagger

Περίληψη

This paper presents the Mnemosyne tagger, a tool for dealing with ambiguity in computing environments. When the POS of a word is ambiguous, the computer must consider all its possible syntactic roles and, if more than one rule is activated, to produce all the structures they dictate. The Mnemosyne tagger is aimed at removing the lexical ambiguity in Modern Greek. It is based not on statistics but on the corresponding linguistic environment of the words. It supports the complete removal of lexical ambiguity not only to remove ambiguity in the POS but also in the gender and in the case of the ambiguous word. The tagger characterizes words that do not exist in the computational lexicon.