Oriya Multiword Chunking using Lexical knowledge base of verbs

The multiword chunking is otherwise thought of as the shallow parsing technique which identifies the multiword chunks and their interdependencies. The paper presents the proposed solution to the problem. Here we have designed the model of the proposed syntactic processor which uses lexical knowledge base of verbs for identifying intra chunk boundaries and finally forming the inter dependencies graph. The lexical knowledge base of verbs is used as there is basic assumption that verbs determine the semantics of a sentence and its surface realization is based on this fact. Verbs impose syntactic restriction over the grammatical representation of the sentence. Further NLP tasks are based on the relations that verbs established with the other words. Several hands on experiments has been conducted to establish the fact that the performance of present scheme exceeds the rule based scheme Key w ords: MWC analyzer, inflection, Tagging, Dependency graphs, free word order, semantic role labeling