TEXT MINING FROM BIOMEDICAL DOMAIN USING A FULL PARSER

Text mining is vital for knowledge cultivation, keeping this in perspective we have focused on developing a system which uses a full parser for analyzing the text, grammar towards the biomedical arena. We proposed a preprocessor to overcome the shortcomings of full parsing and modules to handle the partial outcome. The developed system, not only has the viability to be maintained easily, but also can adapt itself for a particular domain. In the primary experiment, out of 131 argument structures extracted from 96 sentences, 32 were extractable, 33 with ambiguity and the remaining 66 (non-extractable) for which partial result was determined. The work produced better result than the other full parser with reduced count of failure in extraction and ambiguity.