Is Channel Independent strategy optimal for Time Series Forecasting?

TL;DR

This paper introduces the CSC and CR strategies to improve channel independent time series forecasting, demonstrating enhanced performance and efficiency over existing methods, and questions the traditional reliance on static channel strategies.

Contribution

It proposes the CSC and CR methods that enhance channel independent strategies, offering better performance and efficiency, and explores the optimality of using same-channel historical data for forecasting.

Findings

CSC improves CI performance and reduces parameters by over 10 times.

CR achieves competitive results with deep models.

Challenging the assumption that same-channel historical data is always best for forecasting.

Abstract

There has been an emergence of various models for long-term time series forecasting. Recent studies have demonstrated that a single linear layer, using Channel Dependent (CD) or Channel Independent (CI) modeling, can even outperform a large number of sophisticated models. However, current research primarily considers CD and CI as two complementary yet mutually exclusive approaches, unable to harness these two extremes simultaneously. And it is also a challenging issue that both CD and CI are static strategies that cannot be determined to be optimal for a specific dataset without extensive experiments. In this paper, we reconsider whether the current CI strategy is the best solution for time series forecasting. First, we propose a simple yet effective strategy called CSC, which stands for $\mathbf{C}$hannel $\mathbf{S}$elf-$\mathbf{C}$lustering strategy, for linear models. Our Channel Self-Clustering (CSC) enhances CI strategy's performance improvements while reducing parameter size, for exmpale by over 10 times on electricity dataset, and significantly cutting training time. Second, we further propose Channel Rearrangement (CR), a method for deep models inspired by the self-clustering. CR attains competitive performance against baselines. Finally, we also discuss whether it is best to forecast the future values using the historical values of the same channel as inputs. We hope our findings and methods could inspire new solutions beyond CD/CI.