Authors:
Chiara Francalanci
and
Ajaz Hussain
Affiliation:
Politecnico di Milano, Italy
Keyword(s):
Sentiment Analysis, Semantic Networks, Power Law Graphs, Social Influence.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Biomedical Engineering
;
Business Analytics
;
Data Analytics
;
Data Engineering
;
Data Management and Quality
;
Data Modeling and Visualization
;
Enterprise Information Systems
;
Health Information Systems
;
Information Systems Analysis and Specification
;
Knowledge Discovery and Information Retrieval
;
Knowledge Influence and Influencers
;
Knowledge Management
;
Knowledge-Based Systems
;
Ontologies and the Semantic Web
;
Semantics and Social Media
;
Society, e-Business and e-Government
;
Statistics Exploratory Data Analysis
;
Symbolic Systems
;
Web Information Systems and Technologies
Abstract:
This paper starts from the observation that social networks follow a power-law degree distribution of nodes, with a few hub nodes and a long tail of peripheral nodes. While there exist consolidated approaches supporting the identification and characterization of hub nodes, research on the analysis of the multi-layered distribution of peripheral nodes is limited. In social media, hub nodes represent social influencers. However, the literature provides evidence of the multi-layered structure of influence networks, emphasizing the distinction between influencers and influence. The latter seems to spread following multi-hop paths across nodes in peripheral network layers. This paper proposes a visual approach to the graphical representation and exploration of peripheral layers and clusters to exploit underlying concept of k-shell decomposition analysis. The core concept of our approach is to partition the node set of a graph into hub and peripheral nodes. Then, a power-law based modified
force-directed method is applied to clearly display local multi-layered neighbourhood clusters around hub nodes. Our approach is tested on a large sample of tweets from the tourism domain. Empirical results indicate that peripheral nodes have a greater probability of being retweeted and, thus, play a critical role in determining the influence of content. Our visualization technique helps us highlight peripheral nodes and, thus, seems an interesting tool to the visual analysis of social influence.
(More)