Terminology Management Revisited

摘要

Large repositories publishing and sharing terminological, ontological and linguistic resources are available to support the development and use of translation. However, despite the availability of language resources within online repositories, some natural languages associations cannot be found (rare languages or non-common combinations, etc.). Consequently, multiple tools for composing linguistic and terminological resources offer the possibility to create missing language associations. These generated resources need to be validated in order to be effectively used. Manually checking these resources is a tedious task and in some cases hardly possible due to the large amount of entities and associations to go through or due to the lack of expertise in both languages. To solve this matter and generate sound and safe content, tools are needed to automatically validate and filter associations that make no sense. Hence, a validation tool is based itself on external resources such as parallel corpora which need to be either collected or created and filtered. To solve these matters we propose a set of tools that generate new terminological resources (myTerm) and a filter them using a parallel corpus generated by another tool (myPREP). We describe our methodology for terminology management and we describe its implementation within an original framework.