An Accurate and Fully-Automated Ensemble Model for Weekly Time Series Forecasting

TL;DR

This paper introduces a fully-automated ensemble forecasting model for weekly time series that combines multiple advanced techniques to achieve superior accuracy and reproducibility across diverse datasets.

Contribution

The paper presents a novel stacking ensemble approach using meta-learning with lasso regression to combine four diverse base models for weekly forecasting.

Findings

Outperforms existing benchmarks and state-of-the-art models significantly.

Achieves the most accurate weekly forecasts on the M4 dataset.

Demonstrates robustness across multiple datasets and evaluation metrics.

Abstract

Many businesses and industries require accurate forecasts for weekly time series nowadays. However, the forecasting literature does not currently provide easy-to-use, automatic, reproducible and accurate approaches dedicated to this task. We propose a forecasting method in this domain to fill this gap, leveraging state-of-the-art forecasting techniques, such as forecast combination, meta-learning, and global modelling. We consider different meta-learning architectures, algorithms, and base model pools. Based on all considered model variants, we propose to use a stacking approach with lasso regression which optimally combines the forecasts of four base models: a global Recurrent Neural Network model (RNN), Theta, Trigonometric Box-Cox ARMA Trend Seasonal (TBATS) and Dynamic Harmonic Regression ARIMA (DHR-ARIMA), as it shows the overall best performance across seven experimental weekly datasets on four evaluation metrics. Our proposed method also consistently outperforms a set of benchmarks and state-of-the-art weekly forecasting models by a considerable margin with statistical significance. Our method can produce the most accurate forecasts, in terms of mean sMAPE, for the M4 weekly dataset among all benchmarks and all original competition participants.