Meta-Forecasting by combining Global Deep Representations with Local Adaptation

TL;DR

This paper introduces Meta-GLAR, a meta-learning based forecasting method that combines global deep representations with local adaptation, significantly improving out-of-sample forecasting accuracy for time series.

Contribution

It presents a novel meta-learning approach that learns a closed-form adaptation mechanism for time series forecasting, enhancing out-of-sample performance.

Findings

Meta-GLAR achieves competitive out-of-sample forecasting accuracy.

The method effectively adapts to new time series using learned representations.

Empirical results demonstrate improved robustness over existing methods.

Abstract

While classical time series forecasting considers individual time series in isolation, recent advances based on deep learning showed that jointly learning from a large pool of related time series can boost the forecasting accuracy. However, the accuracy of these methods suffers greatly when modeling out-of-sample time series, significantly limiting their applicability compared to classical forecasting methods. To bridge this gap, we adopt a meta-learning view of the time series forecasting problem. We introduce a novel forecasting method, called Meta Global-Local Auto-Regression (Meta-GLAR), that adapts to each time series by learning in closed-form the mapping from the representations produced by a recurrent neural network (RNN) to one-step-ahead forecasts. Crucially, the parameters ofthe RNN are learned across multiple time series by backpropagating through the closed-form adaptation mechanism. In our extensive empirical evaluation we show that our method is competitive with the state-of-the-art in out-of-sample forecasting accuracy reported in earlier work.