A Corpus-based Study On Terminology Extraction Of Maritime English

Terminology is the key member of the knowledge system, and it reflects the knowledge of one subject. Moreover, it is an effective way to know the general trends of a subject through terminology. The method of terminology extraction is one of the key technologies of building a large-scale ontology automatically or semi-automatically. With the awareness of the vital importance of terminology extraction, many studies have been carried out and great progress has been made. So far, there are mainly three approaches to terminology extraction, and they are linguistic-based, statistical-based and hybrid approaches, which have made a large contribution to the theory of terminology extraction. However, all these terminology extraction approaches are different in efficiency. The present study proposes a new terminology extraction method based on specific domain. The experiment based on Maritime English corpus achieved high precision and recall. Furthermore, the efficiency of the proposed approach is improved, compared to traditional approaches.With the aid of NEC, BNC and Foxpro Programs, the study presents the parameters of evaluating a terminology extraction approach. Moreover, the present study proposes a new terminology extraction approach and establishes a nautical term list. The results show that precision and recall are two important parameters to evaluate the terminology extraction approach. Besides, the proposed approach has higher efficiency in terms of precision and recall than previous study based on frequency. The precision and recall of previous approach is 88.6% and 45.0% respectively, while the precision and recall of the proposed approach is 90.4% and 85.4%, it is clear that the precision and recall have gained 1.8% and 40.4% respectively. Furthermore, the proposed approach is effectively to extract low frequency terms.