Growth and decline of complex networks with time-varying attractiveness

• Main Author: Xie, Xun
• Other Authors: Xiao Gaoxi
• Format: Thesis-Master by Coursework
• Language: English
• Published: Nanyang Technological University 2025
• Subjects: Engineering
• Online Access: https://hdl.handle.net/10356/182608
• Institution: Nanyang Technological University

In the era of increasing number of papers releases, the study of cross-referencing relationships between citations has become more important. In this study, the citation aging effect is researched in network field, and a new model structure is proposed on the basis of the classical aging effect model. Since the original aging effect model only considers that the citation value of a paper decreases with time after its publication, it does not take into account that this citation value first should increases with time and then the probability of its citation likelihood decreases. This study retains the structure of the original aging effect model, with reference to Wang's view [28] on the characteristics of post-publication citations conforming to the lognormal distribution. Specifically, we propose the idea of using the probability density function (PDF) of the lognormal distribution to improve the aging effect. The simulation results show that the new model changes the original structure that only considers the continuous decline of citation probability, and effectively explains the reality that the citation value of a paper increases and declines like the spread of new ideas in social networks. Further, in the new model, there is a location parameter, which can be varied to present two completely different citation situations; and the reasons for such different connections are given. The application significance of the new model lies in the fact that it better fits the actual paper citation network by more accurately reflecting the aging effect, and that the variation of the position parameter can fit citation characteristics of different disciplines, making the new model be generalizable for interdisciplinary research.