GBT: Two-stage transformer framework for non-stationary time series forecasting

This paper shows that time series forecasting Transformer (TSFT) suffers from severe over-fitting problem caused by improper initialization method of unknown decoder inputs, especially when handling non-stationary time series. Based on this observation, we propose GBT, a novel two-stage Transformer framework with Good Beginning. It decouples the prediction process of TSFT into two stages, including Auto-Regression stage and Self-Regression stage to tackle the problem of different statistical properties between input and prediction sequences. Prediction results of Auto-Regression stage serve as a 'Good Beginning', i.e., a better initialization for inputs of Self-Regression stage. We also propose the Error Score Modification module to further enhance the forecasting capability of the SelfRegression stage in GBT. Extensive experiments on seven benchmark datasets demonstrate that GBT outperforms SOTA TSFTs (FEDformer, Pyraformer, ETSformer, etc.) and many other forecasting models (SCINet, N-HiTS, etc.) with only canonical attention and convolution while owning less time and space complexity. It is also general enough to couple with these models to strengthen their forecasting capability. The source code is available at: https://github.com/OrigamiSL/GBT & COPY; 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Automated summary

This paper points out that the time series forecasting Transformer (TSFT) suffers from over-fitting due to improper initialization of unknown decoder inputs, especially for non-stationary time series. To address this issue, the authors propose GBT, a two-stage Transformer framework that decouples the prediction process into Auto-Regression and Self-Regression stages. GBT outperforms state-of-the-art TSFTs and other forecasting models in terms of both accuracy and efficiency.