Influence lifetime: modeling the temporal variation of social influence through domains

• Published in: Social network analysis and mining 2024-08, Vol.14 (1), p.174
• Main Authors: Oufaida, Houda, Lhadj, Lynda Said, Oufaida, Souhila, Grine, Alima, Smaili, Naziha
• Format: Article
• Language: English
• Subjects: Business community; Evolution; Experts; Formalism; Influence; Influencer marketing; Information dissemination; Soccer; Social networks; Topology; Tournaments & championships; Tracking; User behavior; User generated content
• ISSN: 1869-5450; 1869-5469
• DOI: 10.1007/s13278-024-01333-7

Thanks to Social Networks (SNs), influencers have gained a crucial role in the business ecosystem. They are recognized for their substantial impact and credibility, particularly within their domain of expertise. However, tracking their evolution presents challenges due to the complexity and dynamism of social networks. Few research works have recently been proposed to address these issues: Which influencer to choose in a given domain? and how to monitor the evolution of their influence strength over time? In this paper, we propose a model to assess the social influence captured through relevant metrics (activity, topology, and expertise) and to trace the temporal evolution of influence within domains of expertise identified through the LDA topic model. Our contribution is threefold: (1) the dynamic propriety of the SN is handled using the Time Varying Graph TVG formalism (2) the assessment of the influence is computed within each domain of expertise and (3) the influence variability is observed with respect to two dimensions: time and domain. Experiments using two twitter datasets have shown that the expertise score is relevant to assess the influence and that our model traces, for top influencers, an interesting view of the temporal variation of their influence within each domain.