Channel-wise Retrieval for Multivariate Time Series Forecasting
This work improves forecasting accuracy for applications with multivariate time series data, though it is incremental by refining retrieval strategies.
The paper tackled the problem of capturing long-range dependencies in multivariate time series forecasting by proposing CRAFT, a channel-wise retrieval framework that addresses inter-variable heterogeneity, resulting in superior accuracy and practical inference efficiency on seven public benchmarks.
Multivariate time series forecasting often struggles to capture long-range dependencies due to fixed lookback windows. Retrieval-augmented forecasting addresses this by retrieving historical segments from memory, but existing approaches rely on a channel-agnostic strategy that applies the same references to all variables. This neglects inter-variable heterogeneity, where different channels exhibit distinct periodicities and spectral profiles. We propose CRAFT (Channel-wise retrieval-augmented forecasting), a novel framework that performs retrieval independently for each channel. To ensure efficiency, CRAFT adopts a two-stage pipeline: a sparse relation graph constructed in the time domain prunes irrelevant candidates, and spectral similarity in the frequency domain ranks references, emphasizing dominant periodic components while suppressing noise. Experiments on seven public benchmarks demonstrate that CRAFT outperforms state-of-the-art forecasting baselines, achieving superior accuracy with practical inference efficiency.