Certain investigations on word sense disambiguation techniques

Abstract

Word Sense Disambiguation (WSD) is an important problem in the field of Computational Linguistics. It helps to identify the exact meaning for words based on the context in which they appear. There is a lot of ambiguity in human language because many words possess multiple meanings. For example, consider the following two sentences: i)The Board meeting of the company took place yesterday. ii)The teacher wrote on the black board. The meaning of board is different in the two sentences. In the first sentence, board means an organized body of administrators and in the second sentence, board refers to a large vertically positioned flat surface used for writing . WSD is considered as an Artificial Intelligence problem. Word Sense Disambiguation is the process of assigning to every word of a document, the most appropriate meaning (sense) among those mentioned in a lexicon or a thesaurus. WSD plays a vital role in Natural Language Processing and Text Mining tasks, such as machine translation, speech processing, information retrieval and document classification. WSD is a type of classification problem because when a word and its possible dictionary senses are given, the process classifies an occurrence of the word in context into one or more of its sense classes. The features of the context such as neighbouring words enable classification to be performed. A number of techniques have been researched ranging from dictionary-based methods that use the knowledge encoded in lexical resources, to supervised machine learning methods where classifiers are trained for words on a corpus of manually annotated senses. Due to several drawbacks in them, new methods are needed to improve the efficiency of WSD. newline