Benchmarking Time Series Foundation Models for Short-Term Household Electricity Load Forecasting

TL;DR

This paper benchmarks recent time series foundation models for short-term household electricity load forecasting, showing they perform comparably to traditional models and excel with larger inputs, while requiring less domain-specific training.

Contribution

It provides a comprehensive comparison of foundation models versus trained-from-scratch approaches for household electricity load forecasting.

Findings

Foundation models perform comparably to TFS Transformer models.

Some foundation models outperform TFS models with larger input sizes.

Foundation models require less domain-specific training and limited data for inference.

Abstract

Accurate household electricity short-term load forecasting (STLF) is key to future and sustainable energy systems. While various studies have analyzed statistical, machine learning, or deep learning approaches for household electricity STLF, recently proposed time series foundation models such as Chronos, TimesFM or Time-MoE promise a new approach for household electricity STLF. These models are trained on a vast amount of time series data and are able to forecast time series without explicit task-specific training (zero-shot learning). In this study, we benchmark the forecasting capabilities of time series foundation models compared to Trained-from-Scratch (TFS) Transformer-based approaches. Our results suggest that foundation models perform comparably to TFS Transformer models, while certain time series foundation models outperform all TFS models when the input size increases. At the same time, they require less effort, as they need no domain-specific training and only limited contextual data for inference.