Time-aware influence minimization via blocking social networks
Publication Type
Conference Proceeding Article
Publication Date
5-2025
Abstract
In this paper, we investigate the Time-aware Influence Minimization (TIMIN) problem in social networks, focusing on minimizing negative influence concerning a critical deadline by temporarily blocking specific nodes in the given social network. First, we introduce the Temporal Linear Threshold (TLT) model, a novel framework that incorporates time delay in influence propagation, the decay of influence power over time, and the lifecycle of influence. Building on this model, we formally define the Timin problem and prove its NP-hardness, monotonicity, and supermodularity. To tackle the Timin problem, we develop the Timin-Greedy, a greedy algorithm that achieves (1 −1/e) approximation. Since exact computation of negative influence spread for any node set in Timin-Greedy is #P-hard, we propose TESTIM, a scalable implementation that provides (1−1/e−ϵ) approximation. To further enhance the efficiency, we introduce NReplacer, a heuristic algorithm leveraging the insight that potential blocking nodes often cluster near the negative source. Our extensive experimental evaluations demonstrate several key findings: (1) TESTIM is up to 10× faster than the baselines while achieving 30%–50% more reductions in negative influence spread, and (2) NReplacer exhibits a 5× speedup compared to TESTIM, with comparable reductions in negative influence spread.
Discipline
Applied Behavior Analysis | Social Psychology
Publication
Proceeding of the 2025 IEEE 41st International Conference on Data Engineering (ICDE), Hong Kong, May 19-23
First Page
557
Last Page
570
Identifier
10.1109/ICDE65448.2025.00048
Publisher
IEEE Computer Society
City or Country
Washington, DC
Citation
CHANG, Xueqin; FU, Jiajie; LIU, Qing; GAO, Yunjun; and ZHENG, Baihua.
Time-aware influence minimization via blocking social networks. (2025). Proceeding of the 2025 IEEE 41st International Conference on Data Engineering (ICDE), Hong Kong, May 19-23. 557-570.
Available at: https://ink.library.smu.edu.sg/sis_research/10387
Additional URL
https://doi.org/10.1109/ICDE65448.2025.00048