Dynamical Graph Echo State Networks with Snapshot Merging for Spreading Process Classification

The Spreading Process Classification (SPC) is a popular application of temporal graph classification. The aim of SPC is to classify different spreading patterns of information or pestilence within a community represented by discrete-time temporal graphs. Recently, a reservoir computing-based model named Dynamical Graph Echo State Network (DynGESN) has been proposed for processing temporal graphs with relatively high effectiveness and low computational costs. Inspired by DynGESN, we propose a novel reservoir computing-based model called the Grouped Dynamical Graph Echo State Network (GDGESN) for dealing with SPC tasks. In this model, a novel augmentation strategy named the snapshot merging strategy is designed for forming new snapshots by merging neighboring snapshots over time, and then multiple reservoir encoders are set for capturing spatiotemporal features from merged snapshots. After those, the logistic regression is adopted for decoding the sum-pooled embeddings into the classification results. Experimental results on six benchmark SPC datasets show that our proposed model has better classification performances than the DynGESN and several kernel-based models.