Disambiguating the ambiguities in natural language processing by using UML models

Pao Yue-kong Library Electronic Theses Database

Disambiguating the ambiguities in natural language processing by using UML models

 

Author: Wang, Lu
Title: Disambiguating the ambiguities in natural language processing by using UML models
Degree: M.A.
Year: 2013
Subject: English language -- Semantics.
Natural language processing (Computer science)
Hong Kong Polytechnic University -- Dissertations
Department: Dept. of English
Pages: vii, 68 pages
Language: English
InnoPac Record: http://library.polyu.edu.hk/record=b2687618
URI: http://theses.lib.polyu.edu.hk/handle/200/7668
Abstract: This research is an analysis of wrong POS tags in Natural Language Processing (NLP) as well as a resolution for some of these wrong tags. It serves three objectives: to identify the wrong tags in the output of the parsing, to analyze the causes underlying these wrong taggings and the characteristics of distribution of wrong tagging in terms of genre and word, and to attempt to provide some resolutions for the wrong tagging by using UML models as well. This study is valuable in that it might help to improve the accuracy of the POS tagging for NLP tool. A Corpus of Travel and Tourism Texts (TnT) is used to obtain data from authentic texts of natural language in English. The TnT corpus contains a collection of eight hundred thousand words and includes four different genres: academic papers, promotional literature, travelogue and online discussion. Each genre contains about two hundred thousand words. After all the texts are tagged with POS tags, a selection of several key words in the texts of each genre are examined closely and analyzed in order to determine their causes before the connection between each genre’s linguistic features and the causes of the incorrect tags is established. The results show that there are three major causes of the wrong tagging: nominal group, preposition ‘to’ and omission. Analysis of the relationship between the genre and these causes reveals that the nominal group is the major contributor in the genre of academic papers, and the omission is the central cause of online discussion. Such characteristics are evidently related to the linguistic features of each genre. Further study of the relationship between the individual key words and the causes of wrong tagging shows that those words which have larger numbers of wrong tagging are lexically ambiguous. On the other hand, words which are not prone to lexical ambiguity are incorrectly tagged mainly because of the statistical approach of the NLP tool.

Files in this item

Files Size Format
b26876188.pdf 452.5Kb PDF
Copyright Undertaking
As a bona fide Library user, I declare that:
  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

     

Quick Search

Browse

More Information