An Automatic Workflow for the Formalization of Scholarly Articles’ Structural and Semantic Elements

摘要: We present a workflow for the automatic transformation of scholarly literature to Linked Open Data (LOD) compliant knowledge base address Task 2 Semantic Publishing Challenge 2016. In this year’s task, we aim extract various contextual information from full-text papers using text mining pipeline that integrates LOD-based Named Entity Recognition (NER) and triplification detected entities. our proposed approach, leverage an existing NER tool ground named entities, such as geographical locations, their LOD resources. Combined with rule-based demonstrate how can both structural (e.g., floats sections) semantic elements authors respective affiliations) provided dataset’s documents. Finally, integrate LODeXporter, flexible exporting module represent results triples in RDF format. As result, generate scalable, TDB-based is interlinked cloud, public SPARQL endpoint task’s queries. Our submission won second place at SemPub2016 challenge average 0.63 F-score.