Enriching Scientific Publications from LOD Repositories Through Word Embeddings Approach

摘要: The era of digitalization is increasingly emphasizing the role Digital Libraries (DL), by increasing requirements and expectations services provided them. interoperability among repositories other resources continues to be a subject research in field. Retrieving publications related particular topic from different DLs, especially diverse domains, require several clicks online visits many points access. However, achieving cross-linking publications, authors data would facilitate scholarly communication general. Starting single point, scholar able find i.e., authors, previously enriched with information repositories. Repositories available as semantic web content, such bibliographic Linked Open Data (LOD) datasets are focus this study. Primarily, we consider existing alignments concepts between Improvements regarding measurements relatedness possible application text-mining techniques. paper introduces preliminary experiments conducted vector space models through TF-IDF Cosine Similarity (CS). Additionally, discusses applying word embedding approach, which focusing mainly on context distributed representations, instead frequency, weighting string matching. We apply contemporary Word2Vec model similar deep learning approach representations.