Metadata complement method by linked open data for literature search

Abstract

In the current papers of the academic society, there are often cases where the literature metadata such as keywords or abstracts are insufficient. In this paper, we propose a new literature retrieval support system that complements the data using the open data and facilitates the search when the relational literature metadata are missing or insufficient using Linked Open Data (LOD). The missing keywords are estimated from the existing titles and abstracts of the literature by extracting the semantic information of the papers. We constructed a prototype of new literature search support system using IEICE Knowledge Discovery System (I-Scover) and confirmed that as a result of experiments using this support system, search accuracy and search time were also improved.