Sylvie Bach (@sylviebach.bsky.social)

Integrating machine learning techniques for critical node identification in complex networks | Scientific Reports In complex networks, the research of finding seed nodes has generated a lot of interest recently due to its significance in numerous applications, such as the advertising field1, power engineering field2, network immunity field3, and viral marketing4. Complex network mining has received a lot of interest5,6. Identifying key nodes is a crucial task in analyzing complex networks. Human mobility and behavioral dynamics plays a crucial role in influencing epidemic spreading processes. Recent comprehensive evaluations highlight that models that incorporate mobility data greatly outperform classic homogeneous-mixing approaches in capturing real-world transmission dynamics7. Methodologically, integrating mobility into epidemic modelling has inspired the deployment of network-based, agent-based, and machine learning techniques that can capture heterogeneous contacts and nonlinear spreading behavior8. Furthermore, empirical data from the pandemic situations highlights the limits of static network centrality measurements by showing that non-pharmaceutical interventions and mobility reductions dynamically change contact patterns and epidemic risk9. Over the period, several centrality measures have been formulated to detect seed nodes. Some of the methods are degree (D)10, Trust PageRank (TPR)11, betweenness (B)12, Katz (K)13, closeness(C)14, PageRank (PR)15, clustering coefficient (CC)16, and some other centralities17,18,19. Local, global, and random-walk measures were the three categories used to classify these centrality methods20. The... https://www.nature.com/articles/s41598-026-40778-y