Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

dual-stream time series forecasting

Dual-stream time series forecasting is a machine learning modeling approach that employs two parallel computational pathways to process historical temporal data and predict future values. In this framework, sequential inputs or their decomposed components, such as trend and seasonal patterns, are routed through two distinct streams tailored to capture complementary dynamics, such as linear and nonlinear behaviors. By utilizing specialized neural network architectures within each branch, the model separates distinct structural characteristics of the time series to better capture underlying temporal dependencies before recombining the representations into a unified forecast.

1 item

xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition

xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition

Artyom Stitsyuk, Jaesik Choi

OrganizationsINEEJIKorea Advanced Institute of Science and Technology

Why you should read this

Proposes xPatch, an effective non-transformer architecture combining exponential moving average seasonal-trend decomposition with dual MLP and CNN streams to surpass attention-based models in long-term time series forecasting.

In recent years, the application of transformer-based models in time-series forecasting has received significant attention. While often demonstrating promising results, the transformer architecture encounters challenges in fully exploiting the temporal relations within time series data due to its attention mechanism. In this work, we design eXponential Patch (xPatch for short), a novel dual-stream architecture that utilizes exponential decomposition. Inspired by the classical exponential smoothing approaches, xPatch introduces the innovative seasonal-trend exponential decomposition module. Additionally, we propose a dual-flow architecture that consists of an MLP-based linear stream and a CNN-based non-linear stream. This model investigates the benefits of employing patching and channel-independence techniques within a non-transformer model. Finally, we develop a robust arctangent loss function and a sigmoid learning rate adjustment scheme, which prevent overfitting and boost forecasting performance.

Added

2026-09-26