The Impact of Temporal Evolution on the Decay of Representation Effectiveness for Research Topics across Disciplines in Heterogeneous Knowledge Networks
The Impact of Temporal Evolution on the Decay of Representation Effectiveness for Research Topics across Disciplines in Heterogeneous Knowledge Networks
Author(s)
Zheng, Yang
Wang, Xiaoguang
Xu, Zhongxiang
Wang, Hongyu
Issue Date
2026-03-12
Keyword(s)
Knowledge networks
Representation effectiveness
Research topics
Hotness prediction
Scientometrics
Abstract
Research topics are key to understanding disciplinary evolution, yet existing studies rarely explain why topic-hotness prediction accuracy decays over time. This study examines temporal decay in research topic representations within heterogeneous knowledge networks and compares decay patterns across the social sciences, applied sciences, and natural sciences. Using author keywords as topic proxies, we integrate semantic features, citation structure, and historical keyword-frequency sequences in a multi-graph representation learning framework, evaluated by MSE, MAE, and R², and further validated via topic tracking from 2014-2024. Results show a general decay in topic representations, with social sciences remaining stable mid-term then declining, applied sciences decaying fastest, and natural sciences staying relatively stable long-term.
Publisher
iSchools
Series/Report Name or Number
iConference 2026 Proceedings
Type of Resource
Other
Genre of Resource
Conference Poster
Language
eng
Permalink
https://hdl.handle.net/2142/132971
Copyright and License Information
Copyright 2026 is held by Yang Zheng, Xiaoguang Wang, Zhongxiang Xu, and Hongyu Wang. Copyright permissions, when appropriate, must be obtained directly from the authors.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
