Production of a Prototype Lexical Database Featuring High-speed, High-accuracy Access and Lexical Knowledge Acquisition

点击翻译按钮获取中文摘要

(1) Generation of lexical data : The data from the "Shin-Meikai Kokugo Jiten" (by Sanseido Publishing Co.) have been modified and expanded to generate the lexical data(approx. 170,000 words).(2) Construction of a word lexicon : A partial lexicon containing only the words used in a target domain has been constructed based on a model of the hierarchical structure of information.(3) Basic design of a database management system : The database management system consists of modules for the access, modification, expansion, acquisition, information structure management, and man-machine interface.(4) Implementation of the database management system : The database management system has been implemented on a workstation using the C language.(5) Design and implementation of detailed specifications of lexical data : Detailed specifications of the lexical data have been designed and the above-mentioned lexical data have been processed and implemented on a workstation.(6) Implementation and evaluation of the access module and the knowledge acquisition module : The access module and the knowledge acquisition module of the database management system have been implemented on a workstation using the C language and prolog.(7) Generation of the lexical database : The results of (5) and (6) have been integrated into a lexical database and its performance has been compared with that of a database using conventional access system both in the speed and the accuracy of access, confirming the advantages of the proposed system.(8) Test of usefulness of the lexical database : The proposed lexical database has been tested in the morpheme analysis of newspaper articles and weather reports, and the results have confirmed that the system has achieved the expected speed and accuracy of access as well as the capability of unknown word acquisition, demonstrating the validity of the proposed lexical database.