Persistent Link:
http://hdl.handle.net/10150/297806
Title:
Solving Vocabulary Matching Problems with Wordnet
Author:
Yearsley, Quinten
Issue Date:
2013
Publisher:
The University of Arizona.
Rights:
Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Abstract:
This study investigated the use of the lexical database WordNet to solve vocabulary matching quizzes. By using the relations in WordNet to match words with semantically similar definitions, it is possible to discover current deficiencies in WordNet, and by experimenting with different ways of using WordNet to find matches, some insight was gained into the semantic relations that tend to exist between words and definitions. Several different methods for measuring semantic similarity between words and definitions were tried and compared, including methods using the WordNet hypernym-hyponym hierarchy, methods using glosses, and methods using other relations in WordNet. Two different algorithms for matching given a set of similarity scores were explored: a greedy matching algorithm and a method that searches globally for the match with the maximum score. It was discovered that by scoring matches using paths through different WordNet relations, 82-85% of the words in the quizzes could be correctly matched, and the most useful relations were the hypernym and similar to relations. However, there were many word and definition pairs that were unable to be matched, revealing some places where WordNet could be improved.
Type:
text; Electronic Thesis
Degree Name:
B.S.
Degree Level:
bachelors
Degree Program:
Honors College; Computer Science
Degree Grantor:
University of Arizona
Advisor:
Fong, Sandiway

Full metadata record

DC FieldValue Language
dc.language.isoenen_US
dc.titleSolving Vocabulary Matching Problems with Wordneten_US
dc.creatorYearsley, Quintenen_US
dc.contributor.authorYearsley, Quintenen_US
dc.date.issued2013-
dc.publisherThe University of Arizona.en_US
dc.rightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.en_US
dc.description.abstractThis study investigated the use of the lexical database WordNet to solve vocabulary matching quizzes. By using the relations in WordNet to match words with semantically similar definitions, it is possible to discover current deficiencies in WordNet, and by experimenting with different ways of using WordNet to find matches, some insight was gained into the semantic relations that tend to exist between words and definitions. Several different methods for measuring semantic similarity between words and definitions were tried and compared, including methods using the WordNet hypernym-hyponym hierarchy, methods using glosses, and methods using other relations in WordNet. Two different algorithms for matching given a set of similarity scores were explored: a greedy matching algorithm and a method that searches globally for the match with the maximum score. It was discovered that by scoring matches using paths through different WordNet relations, 82-85% of the words in the quizzes could be correctly matched, and the most useful relations were the hypernym and similar to relations. However, there were many word and definition pairs that were unable to be matched, revealing some places where WordNet could be improved.en_US
dc.typetexten_US
dc.typeElectronic Thesisen_US
thesis.degree.nameB.S.en_US
thesis.degree.levelbachelorsen_US
thesis.degree.disciplineHonors Collegeen_US
thesis.degree.disciplineComputer Scienceen_US
thesis.degree.grantorUniversity of Arizonaen_US
dc.contributor.advisorFong, Sandiway-
All Items in UA Campus Repository are protected by copyright, with all rights reserved, unless otherwise indicated.