Univariate vs Multivariate Time Series Forecasting with Transformers

William Michael John Murphy, Ke Chen

Keywords: forecasting, time series, transformers, univariate, multivariate

TL;DR: We achieve SOTA results via a simple method of producing multivariate forecasts in a univariate manner which points to flaws in current architectures.

Abstract: Multivariate time series forecasting is a challenging problem and a number of Transformer-based long-term time series forecasting models have been developed to tackle it. These models, however, are impeded by the additional information available in multivariate forecasting. In this paper we propose a simple univariate setting as an alternative method for producing multivariate forecasts. The univariate model is trained on each individual dimension of the time series. This single model is then used to forecast each dimension of the multivariate forecast in turn. A comparative study shows that our setting outperforms state-of-the-art Transformers in the multivariate setting in benchmark datasets. To investigate why, we set three hypotheses and verify them via an empirical study, which leads to a criterion for when our univariate setting is likely to lead to better performance and reveals flaws in the current multivariate Transformers for long-term time series forecasting.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning