Comparing random walks in graph embedding and link prediction.

Autor: Vital A Jr; Institute of Mathematics and Computer Science, USP, São Carlos, SP, Brazil., Silva FN; The Observatory on Social Media (OSoMe), Indiana University, Bloomington, Indiana, United States of America., Amancio DR; Institute of Mathematics and Computer Science, USP, São Carlos, SP, Brazil.
Jazyk: angličtina
Zdroj: PloS one [PLoS One] 2024 Nov 06; Vol. 19 (11), pp. e0312863. Date of Electronic Publication: 2024 Nov 06 (Print Publication: 2024).
DOI: 10.1371/journal.pone.0312863
Abstrakt: Random walks find extensive applications across various complex network domains, including embedding generation and link prediction. Despite the widespread utilization of random walks, the precise impact of distinct biases on embedding generation from sequence data and their subsequent effects on link prediction remain elusive. We conduct a comparative analysis of several random walk strategies, including the true self-avoiding random walk and the traditional random walk. We also analyze walks biased towards node degree and those with inverse node degree bias. Diverse adaptations of the node2vec algorithm to induce distinct exploratory behaviors were also investigated. Our empirical findings demonstrate that despite the varied behaviors inherent in these embeddings, only slight performance differences manifest in the context of link prediction. This implies the resilient recovery of network structure, regardless of the specific walk heuristic employed to traverse the network. Consequently, the results suggest that data generated from sequences governed by unknown mechanisms can be successfully reconstructed.
Competing Interests: The authors have declared that no competing interests exist.
(Copyright: © 2024 Vital et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)
Databáze: MEDLINE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje