Multivariate Time Series Clustering based on Graph Convolutional Network

Abstract Multivariable time series (MTS) clustering is an important topic in data mining. The major challenge of MTS to capture the temporal correlations and dependencies between multiple variables. In this paper, we propose a novel approach based on graph convolutional network (GCN), which powerful feature extractor for structure data. We regard each variable as node construct edges through correlation Furthermore, GCN deep learning back-ropagation technology are used continuously learn relationship Combining learned variables with characteristics dimensions, comprehensive features can be fused form effective representation task. carry out extensive experimental analysis four open sets six benchmark algorithms, shows superiority proposed method.