English

tsGT: Stochastic Time Series Modeling With Transformer

Machine Learning 2024-04-04 v3

Abstract

Time series methods are of fundamental importance in virtually any field of science that deals with temporally structured data. Recently, there has been a surge of deterministic transformer models with time series-specific architectural biases. In this paper, we go in a different direction by introducing tsGT, a stochastic time series model built on a general-purpose transformer architecture. We focus on using a well-known and theoretically justified rolling window backtesting and evaluation protocol. We show that tsGT outperforms the state-of-the-art models on MAD and RMSE, and surpasses its stochastic peers on QL and CRPS, on four commonly used datasets. We complement these results with a detailed analysis of tsGT's ability to model the data distribution and predict marginal quantile values.

Keywords

Cite

@article{arxiv.2403.05713,
  title  = {tsGT: Stochastic Time Series Modeling With Transformer},
  author = {Łukasz Kuciński and Witold Drzewakowski and Mateusz Olko and Piotr Kozakowski and Łukasz Maziarka and Marta Emilia Nowakowska and Łukasz Kaiser and Piotr Miłoś},
  journal= {arXiv preprint arXiv:2403.05713},
  year   = {2024}
}
R2 v1 2026-06-28T15:14:12.854Z