Abstract

Polysemy in WordNet corresponds to various kinds of linguistic phenomena that can be grouped into five classes. One of them is homonymy that refers to the cases, where the meanings of a term are unrelated, and three of the classes refer to the polysemy cases, where the meanings of a term are related. These three classes are specialization polysemy, metonymy, and metaphoric polysemy.Another polysemy class is the compound noun polysemy. In this thesis, we focus on compound noun polysemy and specialization polysemy. Compound noun Polysemy corresponds to the cases, where we use the modified noun to refer to a compound noun. Specialization polysemy is a type of related polysemy referring to the polysemy cases, when a term is used to refer to either a more general meaning or a more specific meaning. Compound noun polysemy and specialization polysemy in WordNet are considered the main reasons behind the highpolysemous nature of WordNet that make WordNet redundant and too fine grained for natural language processing. Another problem in WordNet is its polysemy representation. WordNet represents the polysemous terms by capturing the different meanings of them at lexical level but without giving emphasis on the polysemy classes these terms belong to. The highpolysemous nature and the polysemy representation in WordNet affect the usability of it as suitable knowledge representation resource for natural language processing applications. In fact, the polysemy problem in WordNet is a challenging problem for natural language processing applications, especially in the field of information retrieval and semantic search. To solve this problem, many approaches have been suggested. Although all the state of the art approaches are good to solve the polysemy problem partially, they do not give a general solution for it. In this thesis, we propose a novel approach to solve the compound noun and specialization polysemy problem in WordNet in the case of nouns. Solving the compound noun polysemy and the specialization polysemy problem is an important step that enhances the usability of WordNet as a knowledge representation resource. The proposed approach is not an alternative to the existing approaches. It is a complementary solution for the state of the art approaches especially the systematic polysemy approaches.