Semantic Extraction in PGF using POS tagged data of Sanskrit

Vol.5 No.2

Year: 2010

Issue: October - December

Title: Semantic Extraction in PGF using POS tagged data of Sanskrit   

Author Name: Smita Selot, Neeta Tripathi , A.S. Zadgaonkar   

Synopsis:   

Extraction of semantics from a text is vital application covering various field of artificial intelligence like natural language processing , knowledge representation , machine learning and so forth . Research work is  being carried for languages in international platform like  English , German , Chinese etc .In India also, research associated with Indian languages like Hindi , Tamil , Bengali and other  regional languages is developing faster . In this paper, we have emphasized upon the use of Sanskrit language for semantic extraction . Sanskrit ,being an order free language with systematic grammar gives an excellent opportunity for extracting semantic with higher efficiency .Panini , an ancient grammarian has introduced six Karakas (cases) to identify the semantic role of word in a sentence . These karkas are analyzed and applied for semantic extraction from the Sanskrit text . Input sentences are  first converted into syntactic structure , these syntactic structures are then used for semantic analysis. The syntactic structures are present in Part of Speech (POS) tagged form and various features of this tagged data is analyzed to extract semantic role of the word in a sentence.  .Finally, semantic label of each word of a  sentence are stored in frames called case frames, which act  as a knowledge representation tool . So mapping of input POS tagged data  to semantic tagged data is done and case frames are  generated . Such system are useful in building question-answer based applications , machine learning , knowledge representation and information retrieval.