Approximating Temporal Networks from Aggregated Network Data

Raji Ghawi, Michael Benzinger, Jurgen Pfeffer

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Temporal network data is typically provided in the form of temporal edge lists, where pairs of connected nodes are associated with their interaction timestamps, enabling construction of network snapshots in accumulative or interval manners. For data privacy and/or data storage purposes, data can be instead available only in an aggregated form with the first- and last interaction times and the frequency. In this paper we address the problem of reconstructing temporal networks from aggregated network data. We propose an approach to derive approximate interaction times through linear interpolation over edge time-spans. Our experiments show the feasibility of constructing approximate temporal networks, where the errors are minimal for longer intervals and/or intervals closer to the boundaries of the system lifespan.

Original languageEnglish
Title of host publicationProceedings - 2023 10th International Conference on Social Networks Analysis, Management and Security, SNAMS 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350318906
DOIs
StatePublished - 2023
Event10th International Conference on Social Networks Analysis, Management and Security, SNAMS 2023 - Abu Dhabi, United Arab Emirates
Duration: 21 Nov 202324 Nov 2023

Publication series

NameProceedings - 2023 10th International Conference on Social Networks Analysis, Management and Security, SNAMS 2023

Conference

Conference10th International Conference on Social Networks Analysis, Management and Security, SNAMS 2023
Country/TerritoryUnited Arab Emirates
CityAbu Dhabi
Period21/11/2324/11/23

Keywords

  • aggregated data
  • interpolation
  • interval compression
  • temporal networks

Fingerprint

Dive into the research topics of 'Approximating Temporal Networks from Aggregated Network Data'. Together they form a unique fingerprint.

Cite this