社交网络中的习惯效应是影响力最大化的潜在因素
Habituation Effect in Social Networks as a Potential Factor Silently Crushing Influence Maximisation Efforts
摘要：
信息传播过程是在真实和数字社交网络中观察到的关键现象。网络成员经常受到来自不同来源的信息的压力，例如提高认识的信息活动、病毒式营销、谣言、假新闻或其他活动的结果。信息经常被重复，这种重复可以以累积影响的形式提高绩效。由于处理信息的能力有限，重复的消息也可能被忽略。学习过程导致重复的信息被忽视，因为它们的内容已经被吸收。在这种情况下，反应性随着重复而降低，并且可以观察到习惯化效应。在这里，我们分析传播过程，同时考虑习惯化效应和性能下降以及接触次数的增加。还考虑了在减少消息数量时的恢复能力。结果表明，即使是低习惯性和传播概率的下降可能会大大影响网络覆盖。这可能导致使用影响最大化方法选择的种子集的潜力显著降低。除了习惯化效应对传播过程的影响外，我们还展示了如何使用顺序播种方法来减少习惯化效应。这表明顺序播种比单阶段播种对习惯化效应不那么敏感，并且它可以用于限制对接收消息过载的用户的负面影响。
英文摘要：
Information spreading processes are a key phenomenon observed within real and digital social networks. Network members are often under pressure from incoming information with different sources, such as informative campaigns for increasing awareness, viral marketing, rumours, fake news, or the results of other activities. Messages are often repeated, and such repetition can improve performance in the form of cumulative influence. Repeated messages may also be ignored due to a limited ability to process information. Learning processes are leading to the repeated messages being ignored, as their content has already been absorbed. In such cases, responsiveness decreases with repetition, and the habituation effect can be observed. Here, we analyse spreading processes while considering the habituation effect and performance drop along with an increased number of contacts. The the ability to recover when reducing the number of messages is also considered. The results show that even low habituation and a decrease in propagation probability may substantially impact network coverage. This can lead to a significant reduction in the potential for a seed set selected with an influence maximisation method. Apart from the impact of the habituation effect on spreading processes, we show how it can be reduced with the use of the sequential seeding approach. This shows that sequential seeding is less sensitive to the habituation effect than single-stage seeding, and that it can be used to limit the negative impact on users overloaded with incoming messages.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
