Word Level Translation (Tamil - English) with word sense disambiguation in Tamil using OntNet

This paper discusses a knowledge-engineering approach for word level translation of Tamil words to English. This approach also shows the need for word sense disambiguation to arrive at the contextual category or meaning of a word when different category/meaning can be assigned to a particular word in different contexts. An ambiguous word in a given Tamil sentence is subjected to contextual analysis. The VPs in the given text are located and then augmented with semantic information. These semantic features are captured using the ontology derived from the sub-categorization features. This semantic information will help in assigning the correct meaning for the given ambiguous word. A rule-based syntactic parser and word sense disambiguator have been developed. This has been tested on Tamil sentences from Tamil newspaper websites and the results are encouraging.