OrderFlowDiT: Diffusion Transformer for Limit-Order Volume Forecasting
In brief
- OrderFlowDiT combines a transformer with UOT loss to model limit-order volumes.
- The model improves transport cost by 45.8% and transport energy score by 26.1% across 30 U.S. stocks.
- Evaluation integrates transport metrics, proper scoring rules, and calibration diagnostics.
Researchers introduce OrderFlowDiT, a diffusion-based transformer model designed to predict limit-order volumes in financial markets. The model incorporates an unbalanced optimal transport loss to better capture price geometry and total volumes, significantly improving accuracy metrics compared to traditional methods.
What it changes for you
- Finance & BankingLimit-order forecasting models could improve accuracy and reduce market risks with this geometric approach.
This summary gives you the essentials. The full article, its figures and quotes are on cs.AI updates on arXiv.org.
Read the original articleAI-generated summary of the source, checked by rules. Report an error