Abstract:

To provide advanced information retrieval and intelligent services using AI and big data processing technologies in R&D sectors, an R&D terminology dictionary that reflects classification information in national R&D fields is required. This study proposes the R&D terminology dictionary construction processes and develops a management system based on these processes. The proposed system supports efficient terminology extraction and cleansing by applying step-by-step processes for establishing R&D terms that reflect research field information using keywords and the Science and Technology Standard Classification of national R&D projects. After automatically extracting and cleansing keywords from the project information, the frequency of each term is calculated. Subsequently, a terminology dictionary is automatically constructed based on these frequencies, and term searching and merging via cleansing criteria functions are utilized for terms subject to manual cleansing. At NTIS(National Science &Technology Information Service), information concerning approximately 57,000 projects is collected yearly. Therefore, the step-by-step terminology dictionary construction and cleansing processes of the R&D terminology dictionary management system proposed in this study can be used to include new terms that occur each year. The proposed system enables users to select the terms related to the desired R&D field of R&D information via the search function of the developed terminology dictionary and perform an extended search using Korean-English terms or related words. As such, the system allows the users to accurately and efficiently find desired information.

Keywords: