Temporal Fusion Transformers for interpretable multi-horizon time series forecasting

Bryan Lim, Sercan Ö. Arık, Nicolas Loeff and Tomas Pfister

International Journal of Forecasting, 2021, vol. 37, issue 4, 1748-1764

Abstract: Multi-horizon forecasting often contains a complex mix of inputs – including static (i.e. time-invariant) covariates, known future inputs, and other exogenous time series that are only observed in the past – without any prior information on how they interact with the target. Several deep learning methods have been proposed, but they are typically ‘black-box’ models that do not shed light on how they use the full range of inputs present in practical scenarios. In this paper, we introduce the Temporal Fusion Transformer (TFT) – a novel attention-based architecture that combines high-performance multi-horizon forecasting with interpretable insights into temporal dynamics. To learn temporal relationships at different scales, TFT uses recurrent layers for local processing and interpretable self-attention layers for long-term dependencies. TFT utilizes specialized components to select relevant features and a series of gating layers to suppress unnecessary components, enabling high performance in a wide range of scenarios. On a variety of real-world datasets, we demonstrate significant performance improvements over existing benchmarks, and highlight three practical interpretability use cases of TFT.

Keywords: Deep learning; Interpretability; Time series; Multi-horizon forecasting; Attention mechanisms; Explainable AI (search for similar items in EconPapers)
Date: 2021
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (66)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0169207021000637
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:intfor:v:37:y:2021:i:4:p:1748-1764

DOI: 10.1016/j.ijforecast.2021.03.012

Access Statistics for this article

International Journal of Forecasting is currently edited by R. J. Hyndman

More articles in International Journal of Forecasting from Elsevier
Bibliographic data for series maintained by Catherine Liu ().