Gravity-Quasi-Laplacian Model Finds Influential Nodes with No Tuning Needed
A parameter-free method that beats 8 benchmarks using degree and k-shell data.
A new paper on arXiv from Shima Esfandiari and Seyed Mostafa Fakhrahmad tackles the classic problem of finding influential nodes in complex networks—a challenge with applications in social network analysis, infrastructure resilience, and information diffusion. Existing methods like k-shell decomposition or centrality measures often suffer from low resolution, parameter dependence, or high computational cost. The authors propose a novel Gravity-Quasi-Laplacian approach that combines a quasi-Laplacian structural measure with a gravity-inspired aggregation process.
The method uses only two simple node attributes—degree and k-shell index—to build a strengthened representation of each node's structural role. It then evaluates local influence through a short-range interaction mechanism with a fixed gravity radius of R=3, making it completely parameter-free and computationally efficient. The approach is interpretable and scales well to large networks, avoiding the tuning overhead of competing algorithms.
In experiments across nine real-world networks, the Gravity-Quasi-Laplacian method outperformed eight state-of-the-art techniques in terms of accuracy, resolution, and simplicity. The results highlight a reliable and scalable tool for identifying key nodes without requiring domain-specific parameter adjustments, a significant step forward for practical network analysis.
- Combines quasi-Laplacian structural measure with gravity-inspired aggregation using only degree and k-shell index.
- Parameter-free: uses a fixed gravity radius of R=3, eliminating manual tuning and reducing computational overhead.
- Outperforms 8 state-of-the-art methods on 9 real-world networks in accuracy, resolution, and simplicity.
Why It Matters
Provides a scalable, tuning-free method to find key influencers in social, communication, and transport networks.