Node Clustering of Time-Varying Graphs Based on Temporal Label Smoothness
Node Clustering of Time-Varying Graphs Based on Temporal Label Smoothness
复制标题
DOI：
--
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
Katsuki Fukumoto;Koki Yamada;Yuichi Tanaka
中科院分区：
文献类型：
--
作者：
Katsuki Fukumoto;Koki Yamada;Yuichi Tanaka
我们基于簇标签随时间平滑变化的假设，提出了一种时变图的节点聚类方法。聚类是机器学习、数据挖掘和信号处理的基本任务之一。尽管大多数现有研究都集中在静态图中节点的聚类，但我们经常遇到时间序列数据的时变图，例如社交网络、大脑功能连接和点云。在本文中，我们将时变图中的节点聚类表述为基于谱聚类的优化问题，并具有节点标签的平滑度约束。对合成图和现实世界时变图进行实验，以验证所提出方法的有效性。
We propose a node clustering method for time-varying graphs based on the assumption that the cluster labels are changed smoothly over time. Clustering is one of the fundamental tasks in machine learning, data mining, and signal processing. Although most existing studies focus on the clustering of nodes in static graphs, we often encounter time-varying graphs for time-series data, e.g., social networks, brain functional connectivity, and point clouds. In this paper, we formulate a clustering of nodes in time-varying graphs as an optimization problem based on spectral clustering, with a smoothness constraint of the node labels. Experiments on synthetic and real-world time-varying graphs are conducted to validate the effectiveness of the proposed approach.