Improving the accuracy of global forecasting models using time series data augmentation

Forecasting models that are trained across sets of many time series, known as Global Forecasting Models (GFM), have shown recently promising results in forecasting competitions and real-world applications, outperforming many state-of-the-art univariate forecasting techniques. In most cases, GFMs are implemented using deep neural networks, and in particular Recurrent Neural Networks (RNN), which require a sufficient amount of time series to estimate their numerous model parameters. However, many time series databases have only a limited number of time series. In this study, we propose a novel, data augmentation based forecasting framework that is capable of improving the baseline accuracy of the GFM models in less data-abundant settings. We use three time series augmentation techniques: GRATIS, moving block bootstrap (MBB), and dynamic time warping barycentric averaging (DBA) to synthetically generate a collection of time series. The knowledge acquired from these augmented time series is then transferred to the original dataset using two different approaches: the pooled approach and the transfer learning approach. When building GFMs, in the pooled approach, we train a model on the augmented time series alongside the original time series dataset, whereas in the transfer learning approach, we adapt a pre trained model to the new dataset. In our evaluation on competition and real-world time series datasets, our proposed variants can significantly improve the baseline accuracy of GFM models and outperform state-of-the-art univariate forecasting methods. (c) 2021 Elsevier Ltd. All rights reserved.

Automated summary

Global Forecasting Models (GFM) trained across multiple time series have shown promising results in forecasting competitions and real-world applications. A novel data-augmentation based framework has been proposed to improve the accuracy of GFM models in less data-abundant settings, using techniques such as GRATIS, MBB, and DBA. The proposed variants significantly outperform state-of-the-art univariate forecasting methods in competition and real-world datasets.