Temporal Patch Shuffle (TPS): Leveraging Patch-Level Shuffling to Boost Generalization and Robustness in Time Series Forecasting

TL;DR

Temporal Patch Shuffle (TPS) is a simple, model-agnostic data augmentation technique for time series forecasting that enhances model generalization and robustness by shuffling overlapping temporal patches while preserving local temporal structure.

Contribution

We introduce TPS, a novel patch-level shuffling augmentation method tailored for forecasting, which improves performance across multiple datasets and models without disrupting temporal coherence.

Findings

TPS consistently improves forecasting accuracy across nine datasets.

The method enhances robustness and generalization in various models.

Ablation studies confirm the effectiveness and design rationale of TPS.

Abstract

Data augmentation is a crucial technique for improving model generalization and robustness, particularly in deep learning models where training data is limited. Although many augmentation methods have been developed for time series classification, most are not directly applicable to time series forecasting due to the need to preserve temporal coherence. In this work, we propose Temporal Patch Shuffle (TPS), a simple and model-agnostic data augmentation method for forecasting that extracts overlapping temporal patches, selectively shuffles a subset of patches using variance-based ordering as a conservative heuristic, and reconstructs the sequence by averaging overlapping regions. This design increases sample diversity while preserving forecast-consistent local temporal structure. We extensively evaluate TPS across nine long-term forecasting datasets using five recent model families (TSMixer, DLinear, PatchTST, TiDE, and LightTS), and across four short-term forecasting datasets using PatchTST, observing consistent performance improvements. Comprehensive ablation studies further demonstrate the effectiveness, robustness, and design rationale of the proposed method.