University of Illinois Urbana-Champaign

Do Information Dissemination Patterns Differ among Platforms with Distinct Mechanisms? A Comparative Study Based on Dynamics of Trending Topics

Yuan, Weikang; Jiang, Zhuoren; Lin, Tianqianjin; Yan, Pengwei; Luo, Siqi; Liu, Xiaozhong

Permalink

https://hdl.handle.net/2142/122784

Description

Title
Do Information Dissemination Patterns Differ among Platforms with Distinct Mechanisms? A Comparative Study Based on Dynamics of Trending Topics
Author(s)
  • Yuan, Weikang
  • Jiang, Zhuoren
  • Lin, Tianqianjin
  • Yan, Pengwei
  • Luo, Siqi
  • Liu, Xiaozhong
Issue Date
2024-03-20
Keyword(s)
  • Multivariate time series prediction
  • Explainable artificial intelligence
  • Dynamic evolution
  • Trending topics
Abstract
Understanding the patterns of information dissemination on social media platforms with different content distribution mechanisms becomes crucial to mitigate potential negative impacts and foster a healthier information ecosystem. However, existing studies still face the limitations of data availability, platform specificity, and neglecting concurrent patterns. To address these issues, we employ trending topics as an instrument to investigate information dissemination patterns across various platforms. A multivariate time series deep learning model combined with explainable artificial intelligence (XAI) techniques is utilized to capture the relationships and influences among internal popularity and external rivalry in the trending topic dynamics. For comparative study, we've meticulously constructed a large-scale dataset of over 8 million time-series trending topic data from Weibo (emphasis on the social network mechanism) and Douyin (prioritize the recommendation algorithm). We find that dynamic patterns of trending topics are significantly different under distinct platform mechanisms, particularly in the impact of external rivalry. This study introduces a new research perspective to explore information dissemination patterns across platforms with different mechanisms, and provides valuable insights into analysing the potential impact of platform mechanisms on trending topic dynamics.
Publisher
iSchools
Series/Report Name or Number
iConference 2024 Proceedings
Type of Resource
Other
Language
eng
Handle URL
https://hdl.handle.net/2142/122784
Copyright and License Information
Copyright 2024 is held by Weikang Yuan, Zhuoren Jiang, Tianqianjin Lin, Pengwei Yan, Siqi Luo, and Xiaozhong Liu. Copyright permissions, when appropriate, must be obtained directly from the authors.

Owning Collections

iConference 2024 Posters PRIMARY
Posters presented at the 2024 iConference https://www.ischools.org/iconference