Network Estimation via Graphon With Node Features
Network Estimation via Graphon With Node Features
复制标题
DOI：
10.1109/tnse.2020.2973994
复制
发表时间：
2018-09
影响因子：
6.6
通讯作者：
Yi Su;Raymond K. W. Wong;Thomas C.M. Lee
中科院分区：
文献类型：
--
作者：
Yi Su;Raymond K. W. Wong;Thomas C.M. Lee
一种流行的网络分析模型是可交换图形模型(ExGM)，其特征是一种称为图形的二维函数。估计潜在的图形成为此类分析的关键。人们已经提出了几种非参数估计方法，其中一些方法被证明是一致的。然而，如果节点的某些有用特征(例如，社交网络环境中的年龄和学校)是可用的，则这些方法都不是为了结合该信息源来帮助估计而设计的。本文提出了一种融合了邻接矩阵本身和节点特征信息的一致性图解估计方法。我们表明，适当地利用这些特征可以改善估计。提出了一种交叉验证法来自动选择该方法的整定参数。
One popular model for network analysis is the exchangeable graph model (ExGM), which is characterized by a two-dimensional function known as a graphon. Estimating an underlying graphon becomes the key of such analysis. Several nonparametric estimation methods have been proposed, and some are provably consistent. However, if certain useful features of the nodes (e.g., age and schools in a social network context) are available, none of these methods were designed to incorporate this source of information to help with the estimation. This paper develops a consistent graphon estimation method that integrates information from both the adjacency matrix itself and node features. We show that properly leveraging the features can improve the estimation. A cross-validation method is proposed to automatically select the tuning parameter of the method.