CMoS: Rethinking Time Series Prediction Through the Lens of Chunk-wise Spatial Correlations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Si, Haotian, Pei, Changhua, Li, Jianhui, Pei, Dan, Xie, Gaogang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913857980596224
author Si, Haotian
Pei, Changhua
Li, Jianhui
Pei, Dan
Xie, Gaogang
author_facet Si, Haotian
Pei, Changhua
Li, Jianhui
Pei, Dan
Xie, Gaogang
contents Recent advances in lightweight time series forecasting models suggest the inherent simplicity of time series forecasting tasks. In this paper, we present CMoS, a super-lightweight time series forecasting model. Instead of learning the embedding of the shapes, CMoS directly models the spatial correlations between different time series chunks. Additionally, we introduce a Correlation Mixing technique that enables the model to capture diverse spatial correlations with minimal parameters, and an optional Periodicity Injection technique to ensure faster convergence. Despite utilizing as low as 1% of the lightweight model DLinear's parameters count, experimental results demonstrate that CMoS outperforms existing state-of-the-art models across multiple datasets. Furthermore, the learned weights of CMoS exhibit great interpretability, providing practitioners with valuable insights into temporal structures within specific application scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2505_19090
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CMoS: Rethinking Time Series Prediction Through the Lens of Chunk-wise Spatial Correlations
Si, Haotian
Pei, Changhua
Li, Jianhui
Pei, Dan
Xie, Gaogang
Machine Learning
Recent advances in lightweight time series forecasting models suggest the inherent simplicity of time series forecasting tasks. In this paper, we present CMoS, a super-lightweight time series forecasting model. Instead of learning the embedding of the shapes, CMoS directly models the spatial correlations between different time series chunks. Additionally, we introduce a Correlation Mixing technique that enables the model to capture diverse spatial correlations with minimal parameters, and an optional Periodicity Injection technique to ensure faster convergence. Despite utilizing as low as 1% of the lightweight model DLinear's parameters count, experimental results demonstrate that CMoS outperforms existing state-of-the-art models across multiple datasets. Furthermore, the learned weights of CMoS exhibit great interpretability, providing practitioners with valuable insights into temporal structures within specific application scenarios.
title CMoS: Rethinking Time Series Prediction Through the Lens of Chunk-wise Spatial Correlations
topic Machine Learning
url https://arxiv.org/abs/2505.19090