The Building And Application Of Patent Verb Frame Base

Semantic knowledge base is used in many aspects of NLP, such as machine translation,information retrieval, question-answering, etc. To enrich linguistics knowledge and boost thedevelopment of NLP, it is very important to build large scale semantic knowledge base. Fromthe perspectives of computer understanding and natural language processing, the predicateverbs and adjectives act as the center of syntax structure and interpretation of sentence, thebuilding of verb-related semantic knowledge base is the basic and important work forsemantic analysis and machine translation. Based on the needs of promoting internationalexchanges of patents and increasing use of patent resource, this paper is devoted to theresearch on Chinese verbs and the building and application of patent verb frame base.This paper puts forward3Dimensional Dynamic Concept Model (3-DDCM) bycombining philosophical analysis and thinking of materialist dialectics and inspiration ofHowNet biaxial theory of events. According to the3-DDCM, this paper proposes a method ofverb classification through sentence semantic analysis. The method of verb classificationstudies the rough correspondence among the verb classes, HowNet event taxonomy, and verbclasses in Verb Dictionary. This paper puts forward Chunk and Chunk Sentence-BasedConstruction Semantic Relation System, and proposes the sentence meaning expression basedon chunk by absorbing, combining, and reframing HowNet, HNC, Construction Grammar,Linguistics for Knowledge Engineering, etc.According to3-DDCM and Chunk and Chunk Sentence-Based Construction SemanticRelation System, the architecture of verb frame base and the construction of knowledge basesystem are studied and described. A system for automatically building verb frame base isdeveloped and the procedure of building verb frame base is illustrated. Based on the researchresults above, Sentence Trunk is proposed and Semantic Structure Patterns of Sentences isreinterpreted in this paper. Furthermore, a system for chunk and chunk sentence-basedconstruction semantic relation analysis is developed.