Research On Construction Of The Field Semantic Dictionary Based On WordNet And FrameNet

Machine dictionary is an indispensable tool in the course of using computer to deal with natural language. All computational linguistics automatic processing including information extraction word sense disambiguation, information retrievals,machine translation,intelligent answer system,man-computer dialogue, and so on is inseparable from machine dictionary and the information they need comes from machine dictionary as well. Machine dictionary is a knowledge base. Only with the knowledge base language problems can be deal with. Application of machine dictionary is the premise of computational linguistics. The construction of practical effective electronic dictionary always has been a key research problems and technical issues on the natural language processing and intelligent computer field.Chinese and foreign scholars count on numerous discussions on the construction of machine dictionary. In addition WordNet and FrameNet, few machine dictionaries can be used all the time. Most machine dictionaries have a short life. There are two reasons for this phenomenon:one is there's not a consistent view the view on construction of dictionary, the other, some scholars is not thorough about natural language understanding. The construction of machine dictionary has a variety of ways, but few to solve the real problem. The paper reviewed the construction principles for WordNet and FrameNet, explored the understanding mode from the perspective of human cognitive. At last, a set of method of constructing domain semantic dictionary was put forward.In this paper the main work done as follows:(1) A large number of documents are read to sort out the history of dictionary, the relationship between dictionary and other knowledge, including computational linguistics, cognitive science, cognitive linguistics and the lexicography-a relatively independent scientific appearing in the evolution of dictionary.(2) A detail analysis was done on WordNet and FrameNet based on the former literature. There was a detail elaboration about synonym sets and frame semantics. And getting inspiration from them, put forward the new ideas, what's more, built a domain semantic dictionary. After analysis of human cognition and psychology, the author designed the domain semantic model and illustrate the data organization module from the knowledge base to words unit, depending on the theory of thematic role by Gruber, G.Miller synonym set theory, Charles J.Fillmore theory of frame semantics.(3) The dictionary data organization module was given, use the end word's hash table-length index table-words linked list mode, what's more, give a specific storage methods and the specific vocabulary unit block model.(4) The author applied the domain semantic dictionary to the scenario search which was a new search issue. According to the information is provided by role restrictions which belong to the specific vocabulary unit, the ambiguity issue is expected to solve. The scenario search utilizes matching algorithms search through the relevant terms of the static and dynamic properties, set the threshold value, indicating the final result vocabulary.