FilterTS: Comprehensive Frequency Filtering for Multivariate Time Series Forecasting

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Yulong, Liu, Yushuo, Duan, Xiaoyi, Wang, Kai
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915276112527360
author Wang, Yulong
Liu, Yushuo
Duan, Xiaoyi
Wang, Kai
author_facet Wang, Yulong
Liu, Yushuo
Duan, Xiaoyi
Wang, Kai
contents Multivariate time series forecasting is crucial across various industries, where accurate extraction of complex periodic and trend components can significantly enhance prediction performance. However, existing models often struggle to capture these intricate patterns. To address these challenges, we propose FilterTS, a novel forecasting model that utilizes specialized filtering techniques based on the frequency domain. FilterTS introduces a Dynamic Cross-Variable Filtering Module, a key innovation that dynamically leverages other variables as filters to extract and reinforce shared variable frequency components across variables in multivariate time series. Additionally, a Static Global Filtering Module captures stable frequency components, identified throughout the entire training set. Moreover, the model is built in the frequency domain, converting time-domain convolutions into frequency-domain multiplicative operations to enhance computational efficiency. Extensive experimental results on eight real-world datasets have demonstrated that FilterTS significantly outperforms existing methods in terms of prediction accuracy and computational efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2505_04158
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FilterTS: Comprehensive Frequency Filtering for Multivariate Time Series Forecasting
Wang, Yulong
Liu, Yushuo
Duan, Xiaoyi
Wang, Kai
Machine Learning
Multivariate time series forecasting is crucial across various industries, where accurate extraction of complex periodic and trend components can significantly enhance prediction performance. However, existing models often struggle to capture these intricate patterns. To address these challenges, we propose FilterTS, a novel forecasting model that utilizes specialized filtering techniques based on the frequency domain. FilterTS introduces a Dynamic Cross-Variable Filtering Module, a key innovation that dynamically leverages other variables as filters to extract and reinforce shared variable frequency components across variables in multivariate time series. Additionally, a Static Global Filtering Module captures stable frequency components, identified throughout the entire training set. Moreover, the model is built in the frequency domain, converting time-domain convolutions into frequency-domain multiplicative operations to enhance computational efficiency. Extensive experimental results on eight real-world datasets have demonstrated that FilterTS significantly outperforms existing methods in terms of prediction accuracy and computational efficiency.
title FilterTS: Comprehensive Frequency Filtering for Multivariate Time Series Forecasting
topic Machine Learning
url https://arxiv.org/abs/2505.04158