Information Extraction Model based on Semantic Role and Conceptual Graph

Because lack of semantic information in traditional information extraction, the accuracy is not high in extraction. In order to solve the problem, this article proposes a novel method in information extraction which is based on Semantic role and conceptual graphs. On one hand, the process transforms the shallow semantic information into conceptual graphs, and accurately describes the main meaning of sentence. On the other hand, the calculator computes the similarity of conceptual graphs so as to distinguish the different domains of information. Meanwhile, the mapping rules will be obtained by using semantic role to improve the quality of extraction, this result to 80% F-measure.