Abstract
Current time series forecasting methods can be broadly classified into twocategories: Channel Independent (CI) and Channel Dependent (CD) strategies,both aiming to capture the complex dependencies within time series data.However, the CI strategy fails to exploit highly correlated covariateinformation, while the CD strategy integrates all dependencies, includingirrelevant or noisy ones, thus compromising generalization. To mitigate theseissues, recent works have introduced the Channel Clustering (CC) strategy bygrouping channels with similar characteristics and applying different modelingtechniques to each cluster. However, coarse-grained clustering cannot flexiblycapture complex, time-varying interactions. Addressing the above challenges, wepropose TimeFilter, a graph-based framework for adaptive and fine-graineddependency modeling. Specifically, after constructing the graph with the inputsequence, TimeFilter filters out irrelevant correlations and preserves the mostcritical ones through patch-specific filtering. Extensive experiments on 13real-world datasets from various application domains demonstrate thestate-of-the-art performance of TimeFilter. The code is available athttps://github.com/TROUBADOUR000/TimeFilter.