arXiv：cs.LG· Guilherme Afonso Galindo Padilha, Paulo Salgado Gomes de Mattos Neto, Rafael Menelau Oliveira e Cruz·· 4 小时前AI 评分45
我们真的在评测预测模型吗？预处理对时间序列性能的影响
Are We Really Benchmarking Forecasting Models? The Impact of Preprocessing on Time Series Performance
AI 导读
一项预处理感知基准在 29,000 条 M4 时间序列上，用 16 种可逆预处理流程评测 11 个预测模型，发现逐序列优化预处理可为所有模型带来约 27% 至 87% 的增益。缺乏内置预处理的架构提升最明显，从而能在现代基准中与复杂 SOTA 模型竞争。该基准的全部资源与实验结果已存入一个综合元数据集，以支持后续元学习任务。
正文
Abstract:While established literature underscores the pivotal role of preprocessing in forecasting accuracy, this stage remains largely overlooked in current research. Modern benchmarks typically resort to simple scaling, failing to account for critical transformations required to address nonstationarity, such as differencing. This omission creates a significant structural preprocessing bias that favors models with built-in data treatments while obscuring the true potential of simpler architectures. We study this effect through a preprocessing-aware benchmark that evaluates 11 forecasting models across 16 reversible preprocessing pipelines on 29,000 M4 time series. Our results identify preprocessing as a key driver of forecasting performance. Optimizing preprocessing per series yields gains of approximately 27\% to 87\% across all evaluated models, with architectures lacking internalized preprocessing experiencing the most substantial improvements. This allows simpler architectures to become highly competitive with complex, state-of-the-art models in modern forecasting benchmarks. All resources and experimental results from this benchmark are stored in a comprehensive metadataset to support future metalearning tasks.
| Comments: | 29 pages, 9 figures. Under Review |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.09096 [cs.LG] |
| (or arXiv:2610.09096v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.09096 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Guilherme Afonso Galindo Padilha [view email]
[v1]
Tue, 6 Oct 2026 20:50:28 UTC (1,432 KB)
来源：arXiv：cs.LG · arxiv.org