Semantic Attributes for Citation Relationships: Creation and Visualization

Abstract

This paper presents a method to process a content of research papers in binary PDF format at a server side that gives research information systems new features of citation content analysis. This method efficiently generates JSON versions of PDF documents that allows an easier recognition of papers’ references, in-text references, citation context, etc. As a result, one can parse an extended set of citation data, including a location of citations in a research paper’s structure, frequency of mentioning for the same references, style of reference mentioning and so on. Based on these data we upgrade traditional citation relationships by adding some semantic attributes. Formatting these semantic data according W3C Web Annotation Data Model and integrating the data with some annotation tools, we visualize citation relationships, its semantic attributes and related statistics as annotations for readers of PDF documents from a research information system.