arXiv：cs.LG· Guoxiong Long, Huizhen Huang, Qikun Cai, Tao Huang, Chen Hou·· 4 小时前AI 评分26
差分隐私合成时间序列预测中的序列组装策略评估
Evaluating Sequence Assembly Strategies for Differentially Private Synthetic Time-Series Forecasting
AI 导读
研究考察了差分隐私时间序列生成器产出的固定长度合成窗口在生成后如何组装成长连续序列，并系统变化重叠率与窗口加权方案。在 ETTh1、ETTm1、Weather、Appliances 四类公开数据集和五种预测模型上，TSTR 预测效用由预测模型、重叠率与加权方案共同决定，不同模型有不同组装偏好；提高重叠率通常改善边界连续性，但连续性或单项保真度指标的改善并不稳定降低预测误差。
正文
Abstract:Differentially private time-series generators commonly produce fixed-length synthetic windows, whereas downstream forecasting models often require long continuous training sequences. How these windows are assembled after generation can therefore alter the effective synthetic data presented to a forecaster, even when the trained generator remains unchanged. We study this post-generation sequence assembly process by systematically varying overlap rates and window-weighting schemes and evaluating the resulting sequences in terms of boundary continuity, statistical and temporal fidelity, and Train-on-Synthetic-Test-on-Real (TSTR) forecasting utility. Across four types of public datasets (ETTh1, ETTm1, Weather, and Appliances) and five forecasting models, the results reveal a clear forecaster-dependent assembly principle: downstream TSTR utility is jointly shaped by the forecaster, overlap rate, and window-weighting scheme, leading to distinct assembly preferences across forecasting models. Increased overlap generally improves boundary continuity, but improvements in continuity or individual fidelity diagnostics do not consistently reduce forecasting error, indicating that these diagnostics alone are insufficient for selecting assembly configurations. Complete five-forecaster assembly grids, together with matched Train-on-Real-Test-on-Real (TRTR) references, further characterize these regularities and quantify assembly-dependent utility relative to real-data training. We then validate the identified principles through additional analyses of robustness and generator variability.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.10222 [cs.LG] |
| (or arXiv:2610.10222v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.10222 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Tao Huang [view email]
[v1]
Wed, 7 Oct 2026 15:16:07 UTC (239 KB)
来源：arXiv：cs.LG · arxiv.org