Research on a Combined Ontology-based Text Information Extraction Technology

To improve the effect of the text information extraction(TIE) technology, the paper proposes a new TIE technology named combined ontology-based TIE(COBTIE) which is a fusion of ontology-driven technology and document-driven technology. The proposed technology begins with constructing the COBTIE model based on domain ontology, task ontology and rule ontology; and then, a COBTIE system architecture is put forward to implement the COBTIE model; furthermore, the paper studies some key technologies of the COBTIE system, including how to construct and apply domain ontology, task ontology and rule ontology, how to design and use the evaluation index of the results of TIE, and how to develop the algorithms essential for implementing the system; at last, the COBTIE technology is experimented. Empirical results indicate that COBTIE technology can updates the ontology library using ontology instances and expert knowledge, and can simplify the implementation process of complex TIE tasks. COBTIE technology has the advantages of ontology-driven technology and document-driven technology since it fuses the two technologies, and possesses some good characteristics, such as is easy to realize, has high portability, and has mature technology, et al.