Paper: arXiv 2505.22678

Authors: Jiahao Yang, Ran Fang, Ming Zhang, Jun Zhou

Abstract

In high-frequency trading (HFT), leveraging limit order books (LOB) to model stock price movements is crucial for achieving profitable outcomes. However, this task is challenging due to the high-dimensional and volatile nature of the original data. Even recent deep learning models often struggle to capture price movement patterns effectively, particularly without well-designed features. We observed that raw LOB data exhibits inherent symmetry between the ask and bid sides, and the bid-ask differences demonstrate greater stability and lower complexity compared to the original data. Building on this insight, we propose a novel approach in which leverages the Siamese architecture to enhance the performance of existing deep learning models. The core idea involves processing the ask and bid sides separately using the same module with shared parameters. We applied our Siamese-based methods to several widely used strong baselines and validated their effectiveness using data from 14 military industry stocks in the Chinese A-share market. Furthermore, we integrated multi-head attention (MHA) mechanisms with the Long Short-Term Memory (LSTM) module to investigate its role in modeling stock price movements. Our experiments used raw data and widely used Order Flow Imbalance (OFI) features as input with some strong baseline models. The results show that our method improves the performance of strong baselines in over 75$% of cases, excluding the Multi-Layer Perception (MLP) baseline, which performed poorly and is not considered practical. Furthermore, we found that Multi-Head Attention can enhance model performance, particularly over shorter forecasting horizons.

Complexity vs Empirical Score

  • Math Complexity: 3.5/10
  • Empirical Rigor: 6.5/10
  • Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper applies established deep learning architectures (Siamese networks, LSTM, Multi-Head Attention) with minimal novel mathematical derivations, but rigorously tests on 14 real stocks with specific data (level-II LOB), benchmarks against strong baselines, and reports performance metrics over multiple horizons.

Research Flowchart

  flowchart TD
  A["Research Goal: Improve LOB Prediction for HFT"] --> B["Input Data: Chinese A-Share Market LOB Data"]
  B --> C["Methodology: Siamese Architecture to process Ask/Bid symmetry"]
  C --> D["Enhancement: Integrate Multi-Head Attention with LSTM"]
  D --> E{"Computational Processing"}
  E -->|Raw Data| F["Model Training & Evaluation"]
  E -->|OFI Features| F
  F --> G["Key Findings: >75% Baseline Improvement & MHA Benefit in Short Horizons"]

Cited References (57)

  1. Enhancing stock movement prediction with market index and curriculum learning (2023)
  2. Curriculum learning in deep neural networks for financial forecasting (2019)
  3. A Survey on Curriculum Learning. (2021)
  4. An efficient loss function and deep learning approach for ranking stock returns in the absence of prior knowledge (2024)
  5. Stock selection via spatiotemporal hypergraph attention network: A learning to rank approach (2021)
  6. Essential tensor learning for multimodal information-driven stock movement prediction (2023)
  7. A stock rank prediction method combining industry attributes and price data of stocks (2023)
  8. Enhancing stock ranking forecasting by modeling returns with heteroscedastic Gaussian Distribution (2025)
and 49 more
  1. DR-GAT: Dynamic routing graph attention network for stock recommendation (2024)
  2. Forecasting stock price movement: New evidence from a novel hybrid deep learning model (2022)
  3. Separating the predictable part of returns with CNN-GRU-attention from inputs to predict stock returns (2024)
  4. Forecasting stock price movement: New evidence from a novel hybrid deep learning model (2022)
  5. Stock price prediction based on hybrid CNN-LSTM model (2024)
  6. High Frequency Trading: Overview of Recent Developments (2016)
  7. High-Frequency Trading and its Impact on Market Liquidity: A Review of Literature (2020)
  8. Benchmark Dataset for Short-Term Market Prediction of Limit Order Book in China Markets (2021)
  9. Deep LOB trading: Half a second please! (2022)
  10. High-frequency trading in a limit order book (2008)
  11. A Dynamic Model of the Limit Order Book (2008)
  12. Statistical Modeling of High-Frequency Financial Data (2011)
  13. Tracking Order Book Price Densities with Kalman Filters (2011)
  14. Modelling high-frequency limit order book dynamics with support vector machines (2015)
  15. Analyzing Performance of High Frequency Currency Rates Prediction Model Using Linear Kernel SVR on Historical Data (2017)
  16. Machine learning techniques for price change forecast using the limit order book data (2015)
  17. A Decision Support System for Trading in Apple Futures Market Using Predictions Fusion (2021)
  18. Order book mid-price movement inference by CatBoost classifier from convolutional feature maps (2021)
  19. Mid-price prediction based on machine learning methods with technical and quantitative indicators (2018)
  20. Machine Learning for Forecasting Mid-Price Movements Using Limit Order Book Data (2018)
  21. Universal features of price formation in financial markets: perspectives from deep learning (2018)
  22. Forecasting Stock Prices from the Limit Order Book Using Convolutional Neural Networks (2017)
  23. DeepLOB: Deep Convolutional Neural Networks for Limit Order Books (2019)
  24. Sequence Classification of the Limit Order Book Using Recurrent Neural Networks (2017)
  25. Stock Price Forecasting with Deep Learning: A Comparative Study (2020)
  26. The price impact of order book events (2014)
  27. Deep LOB trading: Half a second please! (2023)
  28. Stock Price Manipulation Detection Using Deep Unsupervised Learning: The Case of Thailand (2021)
  29. Temporal Bag-of-Features Learning for Predicting Mid Price Movements Using High Frequency Limit Order Book Data (2020)
  30. An Optimal Limit Order Book Prediction Analysis Based on Deep Learning and Pigeon-Inspired Optimizer (2021)
  31. Deep order flow imbalance: Extracting alpha at multiple horizons from the limit order book (2023)
  32. Multi-level order-flow imbalance in a limit order book (2018)
  33. The relevance of features to limit order book learning (2022)
  34. Attention is All you Need (2017)
  35. Deep learning for limit order books (2019)
  36. Order imbalance based strategy in high frequency trading (2015)
  37. Multi-level order-flow imbalance in a limit order book (2018)
  38. Are transformers effective for time series forecasting? (2023)
  39. Adam: a method for stochastic optimization (2014)
  40. Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks (2019)
  41. LOBSTER: Limit Order Book Reconstruction System (2011)
  42. Deep attentive survival analysis in limit order books: Estimating fill probabilities with convolutional-transformers (2024)
  43. The short-term predictability of returns in order book markets: A deep learning perspective (2024)
  44. HLOB–Information persistence and structure in limit order books (2025)
  45. Deep Learning with Reinforcement Learning on Order Books. (2024)
  46. Position Attention Mechanism-based Ensemble Network for Trend Prediction of Limit Order Book (2024)
  47. Deep Hawkes process for high-frequency market making (2024)
  48. Lob-based deep learning models for stock price trend prediction: a benchmark study (2024)
  49. A Time Series is Worth 64 Words: Long-term Forecasting with Transformers (2023)

0 of 57 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.