Unlocking the Potential of Deep Learning in Peak-Hour Series Forecasting

TL;DR

This paper introduces Seq2Peak, a framework that significantly improves peak-hour series forecasting by addressing non-stationarity and optimizing peak detection, achieving an average 37.7% performance boost across multiple datasets.

Contribution

The paper proposes Seq2Peak, a novel PHSF framework with CyclicNorm and a peak-hour decoder, effectively bridging the gap between standard TSF and peak-hour forecasting.

Findings

Achieves 37.7% average relative improvement on real datasets.

Effectively mitigates non-stationarity with CyclicNorm.

Enhances peak detection with a trainable, parameter-free decoder.

Abstract

Unlocking the potential of deep learning in Peak-Hour Series Forecasting (PHSF) remains a critical yet underexplored task in various domains. While state-of-the-art deep learning models excel in regular Time Series Forecasting (TSF), they struggle to achieve comparable results in PHSF. This can be attributed to the challenges posed by the high degree of non-stationarity in peak-hour series, which makes direct forecasting more difficult than standard TSF. Additionally, manually extracting the maximum value from regular forecasting results leads to suboptimal performance due to models minimizing the mean deficit. To address these issues, this paper presents Seq2Peak, a novel framework designed specifically for PHSF tasks, bridging the performance gap observed in TSF models. Seq2Peak offers two key components: the CyclicNorm pipeline to mitigate the non-stationarity issue and a simple yet effective trainable-parameter-free peak-hour decoder with a hybrid loss function that utilizes both the original series and peak-hour series as supervised signals. Extensive experimentation on publicly available time series datasets demonstrates the effectiveness of the proposed framework, yielding a remarkable average relative improvement of 37.7% across four real-world datasets for both transformer- and non-transformer-based TSF models.