Answers are grounded in this paper and cite their sources.
Influence maximisation in social networks seeks to identify users who can trigger large-scale information cascades. Many seed node selection approaches are limited as they neglect both global and local structural factors. To address this, we propose the Influence Driven Decision (IDD) Score, a novel metric that integrates global criteria (shares, likes, and trending score) with local features (out-degree) to capture user influence more effectively. The model defines edge weights as the average activity of connected nodes and selects seed nodes based on the highest IDD scores. To further enhance propagation, we introduce a new link-generation strategy. Using Tarjan's Algorithm, weak links that constrain diffusion are identified and replaced with stronger neighbours based on preferential attachment, thereby reinforcing network connectivity and improving the spread potential. Experiments on real-world datasets, including Twitter and Facebook, demonstrate that the proposed IDD-ICM model significantly outperforms traditional centrality-based and diffusion models. Results consistently show higher diffusion coverage, with improvements reported in percentage points (pp), confirming the effectiveness of IDD-ICM in influence maximisation. On Twitter, adoption improvements range from +8 pp (over LTM) to +28 pp (over SIR), while on Facebook, improvements range from +5 pp (over LTM) to +24 pp (over SIR).
No takes yet. Share an insight, caveat, or question.
Susi et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: