分层补丁机制让时间序列预测更准，同时捕捉周期与残差特征。
HPMixer: Hierarchical Patching for Multivariate Time Series Forecasting
- 分两路建模：周期性用可学习循环模块，残差用小波变换提取频域特征。
- 在多个基准数据集上性能超越或媲美最新模型，长期预测效果显著。
- 适合需要高精度多变量时间序列预测的研究者和工业应用。
在长期多变量时间序列预测中，有效捕捉周期模式和残差动态至关重要。为在标准深度学习基准设置下解决这一问题，我们提出分层补丁混合器（HPMixer），以解耦但互补的方式建模周期性和残差。周期成分采用可学习循环模块[7]，并结合非线性通道级MLP提升表达能力；残差成分通过可学习平稳小波变换（LSWT）提取稳定、平移不变的频域表示。随后，通道混合编码器建模显式跨通道依赖，两级非重叠分层补丁机制捕捉粗粒度与细粒度残差变化。通过解耦周期建模与结构化多尺度残差学习的结合，HPMixer构建了高效框架。在多个标准多变量基准上的大量实验表明，该方法性能达到或超过近期基线。
原文摘要 · Abstract (English)
In long-term multivariate time series forecasting, effectively capturing both periodic patterns and residual dynamics is essential. To address this within standard deep learning benchmark settings, we propose the Hierarchical Patching Mixer (HPMixer), which models periodicity and residuals in a decoupled yet complementary manner. The periodic component utilizes a learnable cycle module [7] enhanced with a nonlinear channel-wise MLP for greater expressiveness. The residual component is processed through a Learnable Stationary Wavelet Transform (LSWT) to extract stable, shift-invariant frequency-domain representations. Subsequently, a channel-mixing encoder models explicit inter-channel dependencies, while a two-level non-overlapping hierarchical patching mechanism captures coarse- and fine-scale residual variations. By integrating decoupled periodicity modeling with structured, multi-scale residual learning, HPMixer provides an effective framework. Extensive experiments on standard multivariate benchmarks demonstrate that HPMixer achieves competitive or state-of-the-art performance compared to recent baselines.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。