Positional analysis in cross-media information diffusion networks

• DOI: 10.1007/s41109-018-0108-x
• Publication Date: 2019-01-15T13:02:25Z
• Authors: Tobias Hecking; Laura Steinert; Victor H. Masias; H. Ulrich Hoppe
• Keywords: Hyperlink; Non-negative Matrix Factorization

This paper describes a network reduction technique to reveal possibly hidden relational patterns in information diffusion networks of interlinked content published across different types online media. Topic specific items such as tweets (Twitter), web pages, or versions Wikipedia articles can reference each other through hyperlinks, revisions, retweet relationships, and thus, constitute that reflects the dissemination on web. Beyond focusing structural linking alone, temporal aspect is explicitly taken into account by modelling edge weight between two according difference their publication times. Non-negative matrix factorisation (NMF) applied decompose resulting groups nodes occupying similar positions, which means they have abilities spread receive from nodes. allows for an easier observation basic underlying structure cross-media main pathways. The utility approach differences techniques will be demonstrated along application scenarios related popular news stories media 2016.