arXiv：cs.LG· Bingqi Lian, Linfeng Cheng, Mei Lu, Jerry Wu·· 3 小时前AI 评分35
信号简化不等于预测简化：诊断短周期波动率预测中的残差神经网络
Signal Simplification Is Not Predictive Simplification: Diagnosing Residual Neural Forecasting in Short-Horizon Volatility
AI 导读
一项针对五类高流动性美国资产的短周期波动率预测研究发现，统计第一阶段成功并不保证残差目标更易学习。HAR 类模型优于 AR、MA 和 ARIMA，其拟合残差方差比目标低约 82%，滚动样本外 HAR 误差方差降低约 74%，但残差 LSTM 增强反而使平均 MSE 从 0.3049 升至 0.3594，且每类资产均恶化。
正文
Abstract:Hybrid statistical-neural pipelines often assume that a successful statistical first stage leaves a cleaner and more learnable residual target. We examine that assumption in short-horizon volatility forecasting through a signal-forecast-system diagnostic framework. Across five liquid U.S. assets, a volatility-aligned HAR-style model outperforms AR, MA, and ARIMA. Within expanding training windows, the pre-standardization fitted residual process used to construct residual-LSTM sequences has about 82% lower variance than the corresponding target and near-zero lag-1 autocorrelation; independently, rolling pseudo-out-of-sample HAR errors show about 74% variance reduction and similarly weak lag-1 dependence. Residual-only LSTM augmentation nevertheless raises mean squared error from 0.3049 to 0.3594 on average, with deterioration on every asset. Pure LSTM records the lowest selected pseudo-out-of-sample MSE, 0.2649, while the residual hybrid requires substantially more end-to-end runtime without improving accuracy. We describe this pattern as forecaster-preconditioner asymmetry: first-stage forecasting success and statistical residual simplification need not translate into useful downstream neural preconditioning.
| Comments: | Accepted at the 10th Computational Methods in Systems and Software (CoMeSySo 2026). 17 pages, 4 figures |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.03019 [cs.LG] |
| (or arXiv:2610.03019v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03019 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Bingqi Lian [view email]
[v1]
Fri, 2 Oct 2026 08:55:04 UTC (484 KB)
来源：arXiv：cs.LG · arxiv.org