2020/08/29 by Athanasios N. Nikolakopoulos, Nikolakopoulos, Athanasios N.
Computer Science · Mathematics · #FOS: Computer and information sciences #Information Retrieval (cs.IR) #Machine Learning (stat.ML) #Social and Information Networks (cs.SI) #cs.IR #cs.SI #stat.ML
paper · pdf · doi:10.48550/arxiv.2008.12916
34 pages, 11 figures; Preliminary versions of (part of) this work have been presented as an ACM WSDM conference paper (https://dl.acm.org/doi/10.1145/2433396.2433415), as well as in arXiv:1506.00092 (This article supersedes arXiv:1506.00092) v2: corrected numerical example
arxiv created 2020/09/01 · arxiv updated 2020/09/02
We revisit the Random Surfer model, focusing on its--often overlooked--Teleportation component, and we introduce NCDawareRank; a novel ranking framework designed to exploit network meta-information as well as aspects of its higher-order structural organization in a way that preserves the mathematical structure and the attractive computational characteristics of PageRank. A rigorous theoretical exploration of the proposed model reveals a wealth of mathematical properties that entail tangible benefits in terms of robustness, computability, as well as modeling flexibility and expressiveness. A set of experiments on real-work networks verify the theoretically predicted properties of NCDawareRank, and showcase its effectiveness as a network centrality measure.