XML based language resource management

Abstract

Abstract� this paper introduces the language resource ar ticle structure using XML (e;densible markup Language) and methodology for effective language resource management.Researches for the Natural Language Processing were just focused on information retrieval, analysis for lexical analysis, specific data extraction(speech recognition and document summarization etc ... ), not on effective language resource management area. In a ddition, though computational linguis tics researchers tried to shrink difference between computer science and linguistics, their efforts were focused on the lexical analysis and making computer oriented linguistic roles. We see langua ge processing in different point. If there were sufficient computerized language resources and analysis tools, NaturalLanguage Processing progressed much more. In this paper, we will introduce the XML based document structure for lan guage resource management lind web based language resource management graphic user interfaces which are designed to support language resource purifying and management. The XML document structure for language resource is designed to obtain multiple articles in onc XML document. It aids multi lingual translation system and terminology search system. Its structure will gh'e manageability, usability, efficiency for a language resource management. I. Il'I'TRODUCTlONWe will see the method for effective language resource management which is distributed over internet environment using XML [I] and web based application. We can separate information over internet with following types: text based data, multimedia related data, and general binary data I.Except text based data, other data type have their own meta information which describes their own identity and special information: an example, the name of composer, total playing time, compress ratio, copyright and distributor's name are stored in audio file. Similarly, the camera brand name, ISO value and exposure value for a digital picture file.These data make easier data categorization, management, searching. However, text based data are hardest data type in searching, management, translation though text based data is oldest computerized infonnation type. We can think that it is due to variable usage and format in text type data: HTML, PDF (Portable Document Format by Adobe corp.), DOC (Microsoft Word). These variety and process hardness makes Natural Language Process as major part of Artifi cial Intelligence and Computer Science.Recently, Researches for XML shows many possibilities '. Generally, Binary data means compressed files, executable application