arXiv cs.AI· Jung Hun Phee, Tucker Balch·· 12 小时前AI 评分20
OrderFlowDiT：用传输正则化扩散 Transformer 预测限价单供给
OrderFlowDiT: Transport-Regularized Diffusion Transformer for Probabilistic Limit-Order Supply Forecasting
AI 导读
OrderFlowDiT 是一个面向未归一化买卖盘量价场的条件扩散 Transformer，通过引入非平衡最优传输（UOT）辅助损失，在 30 只美股 1950 个验证条件下将全场传输成本降低 45.8%、传输能量分数降低 26.1%，且每只股票均有改善。研究同时提出结合价格位移与总量敏感性的全场评估框架，将场精度与预测不确定性一并衡量。
正文 · 原文
Abstract:We study the conditional distribution of limit-order volume submitted over the next five minutes, given the current order book and recent order flow. The task requires capturing price placement, total volume, and predictive uncertainty jointly. Coordinate-wise diffusion losses do not explicitly account for price displacement, while normalizing fields to unit mass removes information about total volume. Agreement on selected summary statistics also does not establish fidelity to the joint field distribution. We introduce OrderFlowDiT, a conditional Diffusion Transformer for unnormalized bid and ask volume-at-price fields, trained with an unbalanced optimal-transport (UOT) auxiliary loss. We also develop a full-field evaluation framework that combines a transport cost sensitive to price placement and total volume with proper scoring rules and calibration diagnostics. In the primary comparison on 1950 validation conditions across 30 U.S. equities, adding UOT reduces full-field transport cost by 45.8% and the transport energy score by 26.1% relative to training with mean squared error alone. Both measures improve for every stock. These findings support incorporating price geometry and volume into generative modeling, with field accuracy and predictive uncertainty evaluated together.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Trading and Market Microstructure (q-fin.TR) |
| Cite as: | arXiv:2610.22727 [cs.LG] |
| (or arXiv:2610.22727v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22727 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jung Hun Phee [view email]
[v1]
Thu, 15 Oct 2026 22:56:40 UTC (217 KB)
来源：arXiv cs.AI · arxiv.org