DGCformer: Deep Graph Clustering Transformer for Multivariate Time Series Forecasting

TL;DR

DGCformer is a novel deep learning model that combines graph convolutional networks and transformers to improve multivariate time series forecasting by effectively grouping relevant variables and applying mixed strategies.

Contribution

It introduces a new approach that integrates graph clustering with transformer-based self-attention for enhanced multivariate time series prediction.

Findings

Outperforms state-of-the-art models on eight datasets.

Effectively groups relevant variables using graph convolutional autoencoder.

Demonstrates the benefit of combining channel-dependent and independent strategies.

Abstract

Multivariate time series forecasting tasks are usually conducted in a channel-dependent (CD) way since it can incorporate more variable-relevant information. However, it may also involve a lot of irrelevant variables, and this even leads to worse performance than the channel-independent (CI) strategy. This paper combines the strengths of both strategies and proposes the Deep Graph Clustering Transformer (DGCformer) for multivariate time series forecasting. Specifically, it first groups these relevant variables by a graph convolutional network integrated with an autoencoder, and a former-latter masked self-attention mechanism is then considered with the CD strategy being applied to each group of variables while the CI one for different groups. Extensive experimental results on eight datasets demonstrate the superiority of our method against state-of-the-art models, and our code will be publicly available upon acceptance.