Resolving Word Sense Ambiguity in Natural Language Processing

Abstract

Word sense disambiguation is a task in which the correct meaning of a given word is determined based upon contextual information. Word sense disambiguation is used in such computer applications as spam identification, text filtering, and language translation. A problem in computer-based approaches to word sense disambiguation is that an information bottleneck can occur when referencing word senses on a large scale. Producing a system that can better the present rates of successful disambiguation would be of considerable benefit in computer applications that deal with parsing, analyzing and interpreting natural language. The primary objectives for this research project were to investigate current computational methods of resolving ambiguity in language, and then to hybridize a select few of these methods to achieve a greater overall rate of successful word sense disambiguation.