IVML  
  about | r&d | publications | courses | people | links
   

G. Drakopoulos, A. Kanavos, Ph. Mylonas, S. Sioutas
Defining and evaluating Twitter influence metrics: a higher-order approach in Neo4j
Social Network Analysis and Mining, Springer, Volume 7, Number 52, December 2017
ABSTRACT
Ranking account influence constitutes an important challenge in social media analysis. Until recently, influence ranking relied solely on the structural properties of the underlying social graph, in particular on connectivity patterns. Currently, there has been a notable shift to the next logical step where network functionality is taken into account, as online social media such as Reddit, Instagram, and Twitter are renowned primarily for their functionality. However, contrary to structural rankings, functional ones are bound to be network-specific since each social platform offers unique interaction possibilities. This article examines seven first-order influence metrics for Twitter, defines a strategy for deriving their higher-order counterparts, and outlines a probabilistic evaluation framework. Experiments with a Twitter subgraph with ground truth influential accounts indicate that a single metric combining structural and functional features outperforms the rest in said framework.
01 December , 2017
G. Drakopoulos, A. Kanavos, Ph. Mylonas, S. Sioutas, "Defining and evaluating Twitter influence metrics: a higher-order approach in Neo4j", Social Network Analysis and Mining, Springer, Volume 7, Number 52, December 2017
[ save PDF] [ BibTex] [ Print] [ Back]

© 00 The Image, Video and Multimedia Systems Laboratory - v1.12