社会网络分析中的不稳定人际影响因素

Dissilient interpersonal influences in social network analysis

本文由科睿研究院杰出成员 Ismat Beg发表于Fuzzy Sets and Systems （sciencedirect.com/science/article/pii/S0165011423000787）

社会网络分析领域的研究人员目前正在研究影响模型。现有文献的基本假设是，两位专家之间的人际影响程度可以在一次会议中确定。第二个假设是，在这次会议中确定的人际影响程度是固定不变的，在随后的会议中不会发生变化。在本文中，我们认为两位或多位专家之间的人际影响程度会随着时间的推移而变化。在第二次或第三次会议上，一个人可能不会发现另一个人同样自信。我们的想法是，可能需要几次会议才能最终确定专家受他人影响的程度。我们为人际影响矩阵定义了一个组成函数，并用它来定义该矩阵在每次会议后的演变过程。我们还说明了人际影响矩阵长期收敛的条件。

Influence models are currently studied by researchers working in the field of social network analysis. The basic assumption in the existing literature is that the degree of interpersonal influence among two experts can be determined in a single meeting. The second presumption is that the degree of interpersonal influence determined in this meeting is stagnant and will not change in the subsequent meetings. In this paper, we assert that degree of interpersonal influence among two or more experts can change over time. One may not find the other person equally assertive in the second or third meeting. The idea is that it may take several meetings to finally be able to declare the degree to which an expert is influenced by others. We define a composition function for the matrix of interpersonal influence and use it to define the evolution process that this matrix goes through after every meeting. We also state the conditions under which the matrix of interpersonal influence converges in the long run. 

全文：

Dissilient interpersonal influences in social network analysis.pdf