Abstract
Computing link analysis algorithm scores for individual nodes in a large-scale graph can be computationally expensive, which may limit practical application. A technique for approximating such a computation may be applied to graphs that can be partitioned into a core subgraph (G1) and a peripheral subgraph (G2) without edges pointing from G1 to G2. The approach may involve pre-computing and caching pruned vectors for nodes within the G1 subgraph. A link analysis algorithm score for a node in the G2 subgraph can then be approximated by calculating a linear combination, such as an arithmetic mean, of the cached vectors corresponding to its G1 neighbors. This technique can shift the problem from being computationally-bound to being more memory-bound, which may be addressed with caching and parallel processing to facilitate scalable, personalized graph ranking.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Kuligin, Leonid, "Scalable Personalized Score Approximation Via Core Subgraph Pre-Computation", Technical Disclosure Commons, (January 29, 2026)
https://www.tdcommons.org/dpubs_series/9239