NPMixer: Hierarchical Neighboring Patch Mixing for Time Series Forecasting

2026-05-08

AI Summary

This work addresses the challenge of jointly modeling local temporal dynamics and cross-variable global dependencies in multivariate time series forecasting by proposing a hierarchical architecture named NPMixer. The method innovatively integrates learnable stationary wavelet transforms with neighboring mixing MLP blocks: the former adaptively decomposes input series into trend and detail components, while the latter captures multi-scale temporal patterns through stacked MLPs over non-overlapping time segments and performs channel mixing on high-frequency components to model inter-variable correlations. Evaluated across 28 experimental settings on seven benchmark datasets, NPMixer achieves the best mean squared error (MSE) performance in 71.4% of cases, significantly outperforming current state-of-the-art models.

Abstract

Multivariate time series forecasting remains a challenge due to the complexity of local temporal dynamics and global dependencies across multiple variables. In this paper, we propose \textbf{N}eighboring \textbf{P}atching \textbf{Mixer} (\textbf{NPMixer}), a hierarchical architecture featuring a Learnable Stationary Wavelet Transform that adaptively learns filter coefficients to decompose signals into trend and detail components in a data-dependent manner. Our framework introduces a Neighboring Mixer Block that captures local temporal dynamics through a series of hierarchical MLP layers operating on non-overlapping patches. Specifically, the mixer block utilizes MLPs to learn temporal patterns within and across these patches, expanding the receptive field to capture multi-scale dependencies. A Channel-Mixing Encoder is applied to high-frequency components to learn channel correlations while preserving the stability of the underlying global trend. Extensive experiments on seven benchmark datasets demonstrate that NPMixer consistently outperforms state-of-the-art models, achieving better performance in 20 out of 28 ($71.4\%$) evaluated experimental setups for MSE.

Problem

• multivariate time series forecasting
• local temporal dynamics
• global dependencies
• time series prediction

Innovation

• Learnable Stationary Wavelet Transform
• Neighboring Mixer Block
• Hierarchical MLP
• Channel-Mixing Encoder
• Multivariate Time Series Forecasting