Improving forecasting by subsampling seasonal time series

TL;DR

This paper introduces a simple subsampling method that constructs multiple seasonal series from original data, improving forecasting accuracy by capturing diverse patterns and combining predictions, especially effective for high-frequency data.

Contribution

The paper proposes a novel subsampling approach to enhance forecasting by leveraging multiple seasonal patterns and combining classical models, reducing reliance on expert model selection.

Findings

Performance improvements on M1, M3, M4 datasets.

Robustness for high-frequency data with multiple seasonalities.

Enhanced forecast accuracy through pattern diversification.

Abstract

Time series forecasting plays an increasingly important role in modern business decisions. In today's data-rich environment, people often aim to choose the optimal forecasting model for their data. However, identifying the optimal model requires professional knowledge and experience, making accurate forecasting a challenging task. To mitigate the importance of model selection, we propose a simple and reliable algorithm to improve the forecasting performance. Specifically, we construct multiple time series with different sub-seasons from the original time series. These derived series highlight different sub-seasonal patterns of the original series, making it possible for the forecasting methods to capture diverse patterns and components of the data. Subsequently, we produce forecasts for these multiple series separately with classical statistical models (ETS or ARIMA). Finally, the forecasts are combined. We evaluate our approach on widely-used forecasting competition data sets (M1, M3, and M4) in terms of both point forecasts and prediction intervals. We observe performance improvements compared with the benchmarks. Our approach is particularly suitable and robust for the data with higher frequency. To demonstrate the practical value of our proposition, we showcase the performance improvements from our approach on hourly load data that exhibit multiple seasonal patterns.