Test-Time Adaptation for Non-stationary Time Series: From Synthetic Regime Shifts to Financial Markets

TL;DR

This paper introduces a test-time adaptation framework for non-stationary time series forecasting and classification, demonstrating its effectiveness on synthetic and real financial data with practical deployment insights.

Contribution

It proposes a lightweight TTA method updating normalization parameters for non-stationary time series, applicable to forecasting and classification tasks, with stability mechanisms.

Findings

Normalization-based TTA improves forecasting on synthetic data.

Batch-normalization updates are robust in financial market scenarios.

Aggressive norm-only adaptation can sometimes degrade performance.

Abstract

Time series encountered in practice are rarely stationary. When the data distribution changes, a forecasting model trained on past observations can lose accuracy. We study a small-footprint test-time adaptation (TTA) framework for causal timeseries forecasting and direction classification. The backbone is frozen, and only normalization affine parameters are updated using recent unlabeled windows. For classification we minimize entropy and enforce temporal consistency; for regression we minimize prediction variance across weak time-preserving augmentations and optionally distill from an EMA teacher. A quadratic drift penalty and an uncertainty triggered fallback keep updates stable. We evaluate this framework in two stages: synthetic regime shifts on ETT benchmarks, and daily equity and FX series (SPY, QQQ, EUR/USD) across pandemic, high-inflation, and recovery regimes. On synthetic gradual drift, normalization-based TTA improves forecasting error, while in financial markets a simple batch-normalization statistics update is a robust default and more aggressive norm-only adaptation can even hurt. Our results provide practical guidance for deploying TTA on non-stationary time series.