Please use this identifier to cite or link to this item:
Title: Application of generic sense classes in word sense disambiguation
Keywords: Word sense disambiguation lexical semantics NLP
Issue Date: 2-Mar-2007
Citation: UPALI SATHYAJITH KOHOMBAN (2007-03-02). Application of generic sense classes in word sense disambiguation. ScholarBank@NUS Repository.
Abstract: This thesis focuses on an approach to address the Knowledge Acquisition Bottleneck in Word Sense Disambiguation. We propose using a generalized set of coarse sense classes at classifier level. Generic nature of the sense classes allows using labeled examples from different words to be used in learning the classes, effectively increasing the amount of available training data. I show, using WordNet lexicographer files as generic classes, that this approach can yield state of the art WSD performance with sparse training data. Further, I argue that the human-created, taxonomy based classes such as WordNet lexicographer files are not ideal for supervised learning, as they are not necessarily coherent with the contextual usage patterns, available for the classifier. I propose using clustering techniques to automatically create generic sense classes that are aimed for better performance of WSD, and show that such classes can improve the WSD performance over manually created classes.
Appears in Collections:Ph.D Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
upali kohomban phdthesis.pdf2.11 MBAdobe PDF



Page view(s)

checked on Dec 9, 2018


checked on Dec 9, 2018

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.