Word Sense Disambiguation (WSD) – a challenge of Natural Language Processing, for Gujarati language. All natural languages have words that mean different thing in different contexts. Human beings are generally good at sensing those ambiguities but it is difficult for computers to understand that. But computers can sense that by following certain algorithms and rules. Four methodologies were discussed in this paper for implementing word sense disambiguation: AI-based method, Knowledge-based method, supervised method and unsupervised method. Seeming that knowledge based method is more accurate, the algorithms that are of knowledge based method, are used for further processing. It includes the standard Lesk algorithm, the simplified lesk algorithm, the lesk algorithm with synonyms and a baseline algorithm. With the comparison of these algorithms it is finally decided to use simplified lesk algorithm and it will be implemented on Java
Keywords
Word Sense Disambiguation, Knowledge-based method, the Lesk Algorithm, the Simplified lesk algorithm
