Study of Methods for Element Data Service for Electronic Documents Related to National R&D Projects

Abstract

Major countries are supporting new knowledge creation and innovative activities by opening data so that the research results carried out by government budgets can serve as public goods. The Republic of Korea has also made research and development(R&D) reports and papers available in electronic file format, which is the result of national R&D programs for the general public. However, the extraction of meaningful information among unstructured data in text format has not satisfied researchers’ expectations. In order to evolve into a customized service reflecting the opinions of researchers, we investigated the demand for necessary contents and services at the planning stage of R&D projects. This study attempts to propose a method to offer significant information which shows a bigger unit than objects based on trend information with extraction and processing demands focusing on R&D reports based on the results of questionnaire survey and interviews with researchers. This study aims to provide the integration service, tentatively named ‘element data service’, of key sentences and table/figure images with a high demand for the utilization of researchers. The main procedure of the proposed method consists of the subject classification of the R&D report, the extraction of table/figure image, and the extraction of main sentence. We used public reports of the same classification published from 2012 to 2016 for the experiment and utilized the extractive summarization method for the copyright protection of report original text. After realizing the simple prototype, we examined the service possibility through the researcher target reviews.