SWIFT: Mapping Sub-series with Wavelet Decomposition Improves Time Series Forecasting

2025-01-27 · Wenxuan Xie, Fanpu Cao

In recent work on time-series prediction, Transformers and even large language models have garnered significant attention due to their strong capabilities in sequence modeling. However, in practical deployments, time-series prediction often requires operation in resource-constrained environments, such as edge devices, which are unable to handle the computational overhead of large models. To address such scenarios, some lightweight models have been proposed, but they exhibit poor performance on non-stationary sequences. In this paper, we propose $\textit{SWIFT}$, a lightweight model that is not only powerful, but also efficient in deployment and inference for Long-term Time Series Forecasting (LTSF). Our model is based on three key points: (i) Utilizing wavelet transform to perform lossless downsampling of time series. (ii) Achieving cross-band information fusion with a learnable filter. (iii) Using only one shared linear layer or one shallow MLP for sub-series' mapping. We conduct comprehensive experiments, and the results show that $\textit{SWIFT}$ achieves state-of-the-art (SOTA) performance on multiple datasets, offering a promising method for edge computing and deployment in this task. Moreover, it is noteworthy that the number of parameters in $\textit{SWIFT-Linear}$ is only 25\% of what it would be with a single-layer linear model for time-domain prediction. Our code is available at https://github.com/LancelotXWX/SWIFT.

📄 PDF Abstract BibTeX arXiv:2501.16178

Code (1)

lancelotxwx/swift 공식 구현 pytorch

Tasks

Edge-computingTime SeriesTime Series Forecasting Time Series Prediction

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Attention 설명 없음

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.

Notes on the SWIFT method based on Shannon Wavelets for Option Pricing -- Revisited

2024-01-03 · Fabien Le Floc'h

This note revisits the SWIFT method based on Shannon wavelets to price European options under models with a known characteristic function in 2023. In particular, it discusses some possible improvements and exposes some c…

AWGformer: Adaptive Wavelet-Guided Transformer for Multi-Resolution Time Series Forecasting

2026-01-28 · Wei Li arxiv

Time series forecasting requires capturing patterns across multiple temporal scales while maintaining computational efficiency. This paper introduces AWGformer, a novel architecture that integrates adaptive wavelet decom…

Computational Efficiency Time Series Forecasting Time Series Prediction

WPMixer: Efficient Multi-Resolution Mixing for Long-Term Time Series Forecasting

2024-12-22 · Md Mahmuddun Nabi Murad, Mehmet Aktukmak, Yasin Yilmaz

Time series forecasting is crucial for various applications, such as weather forecasting, power load forecasting, and financial analysis. In recent studies, MLP-mixer models for time series forecasting have been shown as…

Financial Analysis Load Forecasting Time Series Time Series Forecasting+1

Multilevel Wavelet Decomposition Network for Interpretable Time Series Analysis

2018-06-23 · Jingyuan Wang, Ze Wang, Jianfeng Li, Junjie Wu

Recent years have witnessed the unprecedented rising of time series from almost all kindes of academic and industrial fields. Various types of deep neural network models have been introduced to time series analysis, but …

Deep Learning General Classification Time Series Time Series Analysis+1

SWIFT: Spatio-temporal Wavelet Integrated Forecasting Framework for Workload Traces

2026-05-07 · Zeyuan Ding, Lingfeng Zheng, Dian Ding, Guangtao Xue arxiv

Accurate cloud workload forecasting is pivotal for efficient resource management but remains challenging as workloads are highly volatile and prone to sudden bursts. Although wavelets preserve temporal locality, rigid fi…