Two-stage hybrid models for enhancing forecasting accuracy on heterogeneous time series

TL;DR

This paper introduces a two-stage hybrid modeling approach that combines global and local time series models to improve forecasting accuracy on heterogeneous datasets, addressing challenges of heterogeneity and overfitting.

Contribution

It proposes a novel two-stage framework that first identifies homogeneous patterns with a global model, then applies tailored local models, enhancing forecasting on diverse datasets.

Findings

Significantly outperforms six state-of-the-art models on four datasets.

Effectively captures heterogeneity with a two-stage approach.

Demonstrates robustness across different types of time series data.

Abstract

A time series forecasting model--which is typically built on a single time series--is known as a local time series model (tsLM). In contrast, a forecasting model trained on multiple time series is referred to as a global time series model (tsGM). tsGMs can enhance forecasting accuracy and improve generalisation by learning cross-series information. As such, developing tsGMs has become a prominent research focus within the time series forecasting community. However, the benefits of tsGMs may not always be realised if the given set of time series is heterogeneous. While increasing model complexity can help tsGMs adapt to such a set of data, it can also increase the risk of overfitting and forecasting error. Additionally, the definition of homogeneity remains ambiguous in the literature. To address these challenges, this paper explores how to define data heterogeneity and proposes a two-stage modelling framework: At stage one, a tsGM is learnt to identify homogeneous patterns; and at stage two, tsLMs (e.g., ARIMA) or sub-tsGMs tailored to different groups are learnt to capture the heterogeneity. Numerical experiments on four open datasets demonstrate that the proposed approach significantly outperforms six state-of-the-art models. These results highlight its effectiveness in unlocking the full potential of global forecasting models for heterogeneous datasets.