The Semantic Representation And Implementation Of Chinese FrameNet Based On XML

CFN(Chinese FrameNet) that is a resource of Chinese lexical semantics for the use of computer takes the Fillmore’s frame semantics as the theoretical basis, the FrameNet as the reference, and the Chinese corpus as the evidence. Chinese FrameNet mainly consists of frame library, sentences library and word element library. At present, CFN constructs 323 Chinese frames, involving 3947 word elements, labeled the 20000 sentences, also it provides precious Chinese semantic resources for computer research on Chinese frame semantic analysis and its application.That how to control the Chinese FrameNet knowledge representation of kinds of resource, is the key problem to maximize its application value. In this paper, the extensible markup language (XML) is used to represent the CFN resource data through the use of the unified description method, which makes the computer readable, and suitable for people to read and understand, and used for subsequent research and related semantic auxiliary. The main contents of this paper are as follows:(1) Semantic resources of Chinese FrameNet contains frame library, sentence library and chapter library. Aiming at the structural characteristics of the three libraries, and following the XML specification, according to FrameNet, and the related resources in LTP (Language Technology Platform), it systematically formulated the XML expression system for Chinese FrameNet semantic resource.(2) The frame library, word element library resources are stored in the form of a word document. After learning and mastering the Jacob technology and combining with the component of Microsoft office, it completes the file conversion from a word document to the XML file. At present, the existing resources have been converted, which is a total of 4270 articles.(3) According to the characteristics of labeled sentences and chapter generated XML files, it designs and implements automatic XML generation system for Chinese FrameNet semantic resources sentences and chapter library. At present, it has generated 18000 sentences and 164 chapters of the XML file.(4) In order to allow researchers to more intuitively analyze semantic information resources, aiming at the characteristic of multi angle frame library resources that show the feature of role relationship, with reference to the FrameNet display technology, mastering and using the XSL style sheet make resources can be displayed more beautiful on the web browser. At the same time, constructing index for the unified management of resources enable researchers to retrieve resources more quickly.According to the structural characteristics of Chinese FrameNet resources, it sets up a Chinese FrameNet semantic resources system based on XML language representation. At the same time, it has realized the automatic generation of XML file, achieved the automatic machine efficiently reading, and showed the frame of the XML file in the browser, which directly facilitated the researchers to intuitively understand frame semantic information. Moreover, the system combines with the automatic semantic analysis system, and makes the results to be automatically generated to XML file, which extends the application scope of CFN and lays the foundation for the further related research.