Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting

Abstract

Transformers have shown great power in time series forecasting due to their global-range modeling ability. However, their performance can degenerate terribly on non-stationary real-world data in which the joint distribution changes over time. Previous studies primarily adopt stationarization to attenuate the non-stationarity of original series for better predictability. But the stationarized series deprived of inherent non-stationarity can be less instructive for real-world bursty events forecasting. This problem, termed over-stationarization in this paper, leads Transformers to generate indistinguishable temporal attentions for different series and impedes the predictive capability of deep models. To tackle the dilemma between series predictability and model capability, we propose Non-stationary Transformers as a generic framework with two interdependent modules: Series Stationarization and De-stationary Attention. Concretely, Series Stationarization unifies the statistics of each input and converts the output with restored statistics for better predictability. To address the over-stationarization problem, De-stationary Attention is devised to recover the intrinsic non-stationary information into temporal dependencies by approximating distinguishable attentions learned from raw series. Our Non-stationary Transformers framework consistently boosts mainstream Transformers by a large margin, which reduces MSE by 49.43% on Transformer, 47.34% on Informer, and 46.89% on Reformer, making them the state-of-the-art in time series forecasting. Code is available at this repository: https://github.com/thuml/Nonstationary_Transformers.

Cited by in corpus (11)

• Fredformer: Frequency Debiased Transformer for Time Series Forecasting
• MTSCI: A Conditional Diffusion Model for Multivariate Time Series Consistent Imputation
• MARLP: Time-series Forecasting Control for Agricultural Managed Aquifer Recharge
• Proactive Model Adaptation Against Concept Drift for Online Time Series Forecasting
• Unlocking the Potential of Deep Learning in Peak-Hour Series Forecasting
• DRFormer: Multi-Scale Transformer Utilizing Diverse Receptive Fields for Long Time-Series Forecasting
• MemDA: Forecasting Urban Time Series with Memory-based Drift Adaptation
• Multiscale Representation Enhanced Temporal Flow Fusion Model for Long-Term Workload Forecasting
• Adapformer: Adaptive Channel Management for Multivariate Time Series Forecasting
• Attention for Causal Relationship Discovery from Biological Neural Dynamics
• Blood Glucose Control Via Pre-trained Counterfactual Invertible Neural Networks