Research On Represented And Matched Of The Conceptual Graph Oriented To Semantics Analysis Of The Field Text

Modern social life, information generation, exchange engaged all the time, so that people acquire knowledge economy, culture and life information from a sustainable development. Text as a medium of information resources,and it in the history of human development plays an irreplaceable role, which its analysis can help people to understand and apply knowledge. In addition, a computer as a new automated tools to solve the text analysis in the modern society. We need a good text semantics representation method as a tool to analyze the text semantics. A good representation can not only bring people to remember, understand happiness, while helping to develop computer tools. The existing text semantic analysis focuses on logical reasoning, mathematical calculations and the conclusions to prove. But the observer and analyst use computer to calculate and analyze the text semantics.First of all, this paper introduced from the text representation and semantics. The paper summarizes the common text semantic representation, such as first order predicate logic, semantic role labeling, semantic frameworks and E-A-V(Entity-Attribute-Value) structural concept maps. And then, an example of the text's semantic analysis has given to show its procedures, and the chapter focusing on the content of the concept internal analysis method. Secondly, the concept graph is a proven superior to the classical knowledge representations, which the researcher use it as a semantic representation of Chinese text and semantic analysis to Chinese information processing research to provide new ideas. At the same time, build a conceptual graph based on the results from the syntactic annotation analysis, but the corresponding annotation syntactic analysis and semantic relations relationship is not coordinated. The three basic elements(concepts, Relationships, arcs) and the feature items to extract and recognize by a method is the foundation of the concept knowledge base and analysis of relevant research. Furthermore, Chinese as a character of icons, the internal attribution revealed structural shape. According to E-A-V logic model structure of the concept internal attribution theory applies to the sentence has a main function of the relationship between the structure of a given. But the traditional concept graph applies to a subject-predicate sentence semantic relationship. This paper presents C-A&R(Concept-Attribute & Relationship) graph which in conjunction entity relationships and the concept internal attribution, it as a Chinese sentence semantic representation. Finally, the C-A&R representation as a text indexing combine with the concept of hierarchy and vector space mode VSM to calculate the two C-A&R conceptual graph of semantic similarity. Thesis C-A&R represents and a conceptual graph analysis method has proposed the establishment of a corresponding experimental method to verify the accuracy, effectiveness.