IVE: Enhanced Probabilistic Forecasting of Intraday Volume Ratio with Transformers

2024-11-17

arXiv.org

AI Summary

This paper addresses the low accuracy and high volatility in intraday minute-level volume-weighted average price (VWAP) trading volume ratio forecasting. Methodologically, it proposes a probabilistic Encoder-Decoder Transformer framework: (i) introduces log-normal distribution modeling for volume ratios to enhance numerical stability; (ii) fuses heterogeneous features—including temporal statistics, market liquidity signals, absolute time encoding, and stock-specific attributes; and (iii) employs a greedy-sampling distribution head for end-to-end joint prediction of mean and standard deviation, enabling uncertainty quantification. Contributions include: (i) the first full probabilistic modeling approach for this task; (ii) significant accuracy improvements on highly liquid stocks in Korean and U.S. markets; (iii) consistent outperformance of the VWAP benchmark over a 2.5-month live trading period; and (iv) robust detection of liquidity spikes, thereby supporting low-impact execution decisions.

Technology Category

• Reasoning under Uncertainty: Relational Probabilistic Models
• Machine Learning: Probabilistic Circuits and Graphical Models
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Large-scale security measurements
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

This paper presents a new approach to volume ratio prediction in financial markets, specifically targeting the execution of Volume-Weighted Average Price (VWAP) strategies. Recognizing the importance of accurate volume profile forecasting, our research leverages the Transformer architecture to predict intraday volume ratio at a one-minute scale. We diverge from prior models that use log-transformed volume or turnover rates, instead opting for a prediction model that accounts for the intraday volume ratio's high variability, stabilized via log-normal transformation. Our input data incorporates not only the statistical properties of volume but also external volume-related features, absolute time information, and stock-specific characteristics to enhance prediction accuracy. The model structure includes an encoder-decoder Transformer architecture with a distribution head for greedy sampling, optimizing performance on high-liquidity stocks across both Korean and American markets. We extend the capabilities of our model beyond point prediction by introducing probabilistic forecasting that captures the mean and standard deviation of volume ratios, enabling the anticipation of significant intraday volume spikes. Furthermore, an agent with a simple trading logic demonstrates the practical application of our model through live trading tests in the Korean market, outperforming VWAP benchmarks over a period of two and a half months. Our findings underscore the potential of Transformer-based probabilistic models for volume ratio prediction and pave the way for future research advancements in this domain.

Problem

• Predict intraday volume ratio for VWAP strategies
• Enhance accuracy using Transformer architecture
• Enable probabilistic forecasting of volume spikes

Innovation

• Transformer architecture for intraday volume prediction
• Probabilistic forecasting captures volume ratio variability
• Enhanced VWAP strategies with live trading tests