近期，365英国上市刘凤鸣教授团队在《Expert Systems with Applications》期刊上发表论文“Fractal property: A tool for understanding the generation mechanism of echo chambers”，第一作者为博士生孙英苹，通讯作者为刘凤鸣教授。《Expert Systems with Applications》是运筹学与管理科学领域权威期刊，属于JCR SCI一区期刊，中科院一区Top期刊，近五年影响因子为7.6。

论文简介：

回音室由聚集在社交网络中具有相似观点和情感倾向特征的个体组成，能够加剧偏见和误解，扩大谣言的影响。了解回音室的产生机制对于管理回音室至关重要，而现有研究往往忽视了回音室在其形成的网络中的互动机制。本文研究了回音室是如何通过其形成的网络中的自相似性产生和扩展的。为此，我们识别了回音室及其特征，并通过 BCA-ECN 模型（回音室网络的盒覆盖算法）计算了回音室网络的分形维数。然后，我们进一步提出了重整化的概念以分析回音室网络的自相似性，并探讨回音室网络演化过程中自相似性与回音室变化之间的关系。通过对收集的微博和抖音舆情数据集的分析，我们发现 BCA-ECN 能够准确高效地计算回音室网络的分形维度，而且重整化前后网络拓扑和节点属性特征的近似不变性揭示了回音室网络的自相似性，回音室的出现、增长和稳定与回音室网络的自相似性密切相关。该研究有助于舆论监管部门更好地识别回音室，制定策略来打破回音室网络的自相似性，提高应对效率，优化舆论传播环境。

Abstract:

Echo chambers consisting of individuals gathered in social networks with similar views and emotional tendency characteristics can exacerbate prejudices and misunderstandings and amplify the influence of rumors. Understanding the generation mechanism of echo chambers is essential for managing echo chambers and existing research often ignores the mechanisms by which echo chambers interact in the networks they formed. This paper investigated how echo chambers are generated and extended through self-similarity in the network they formed. To this end, we identified echo chambers and their features and calculated the fractal dimension of the echo chamber network through BCA-ECN model (Box Covering Algorithm for Echo Chamber Network). Then, we further proposed the concept of renormalization to analyse the self-similarity of echo chamber networks as well as to explore the relationship between this self-similarity and echo chamber changes in the evolution of echo chamber networks. By collecting and analysing public opinion datasets of Weibo and Douyin, we found that BCA-ECN can accurately and efficiently compute the fractal dimension of the echo chamber network, and the approximate invariance of the network topology and node attribute characteristics before and after renormalization revealed the self-similarity of the echo chamber network as well as the emergence, growth and stabilization of echo chambers are closely related to the self-similarity of the echo chamber network. The study can help public opinion regulators identify echo chambers better, formulate response strategies to break the self-similarity of echo chamber networks, improve response efficiency, and optimize the public opinion communication environment.

作者简介：

第一作者：孙英苹，博士生，主要从事社交网络与回音室效应研究，目前受国家留学基金委资助在德国慕尼黑工业大学接受博士联合培养。

通讯作者：刘凤鸣，教授，博士生导师，自2008年起在英国365集团公司任职，主要研究领域为信息治理与决策优化，博弈论与量子信息等。发表学术研究论文120余篇，其中在《Knowledge-Based Systems》、《Expert Systems With Applications》、《The Lancet Planetary Health》等期刊发表多篇文章。主持国家社会科学基金项目《大数据支持下网络谣言的智慧治理问题研究》等多项科研项目以及荣获多项学术奖励。担任中国运筹学会智能计算分会第五届理事会副理事长、山东东盟人才产业研究院理事（首席专家）。