xPatch: Dual-Stream Time Series Forecasting with Exponential Seasonal-Trend Decomposition
This work addresses time series forecasting for applications requiring accurate predictions, but it appears incremental as it builds on existing decomposition and patching techniques.
The authors tackled the challenge of transformer-based models not fully exploiting temporal relations in time series forecasting by proposing xPatch, a dual-stream architecture with exponential decomposition, which achieved competitive results on benchmark datasets.
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. The code is available at the following repository: https://github.com/stitsyuk/xPatch.