.png){kind=icon}
.png){kind=icon}
Return
RTGEN plus plus : A Relative Temporal Graph GENerator
.png){kind=icon}
OA
AI
DOI：10.1016/j.future.2023.03.023.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Graph management systems have become popular for storing and querying graph-oriented data, and they are often evaluated with benchmarks based on large-scale graphs. However, obtaining such graphs is difficult due to their limited public availability. Several graph generators have been developed to address this challenge, producing synthetic graphs with characteristics similar to real-world graphs, such as degree distribution, community structure, and diameter. Generating synthetic graphs with constantly changing topology has received less attention despite its importance in developing useful benchmarks for temporal graph systems. In this paper, we present RTGEN++, a temporal graph generator that supports two evolution models. The first model generates temporal graphs by controlling the evolution of their degree distributions, using optimal transport methods to minimize the transformation effort. We also extend our method in order to consider the community structure of the generated graphs. The second model allows one to control the number of added and removed graph entities, thus enabling the modeling of the evolution of real-world graphs in many use cases. Our generator also includes a decorator that adds types and time-varying attributes to nodes and edges, enhancing the generated graphs and aligning with data platforms that use the property graph model. We validate our approach with experiments that demonstrate the reliability of the generated graphs in approximating ground-truth parameters. (c) 2023 Published by Elsevier B.V.
Keywords:
Temporal graphs
Graph generation
Optimal transport
Benchmarking data platforms
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
F
IF:
6.1
Papers:
6.8K
Citations:
2.3W