Paper: arXiv 2503.06929

Authors: Yanlong Wang, Jian Xu, Shao-Lun Huang, Danny Dongning Sun, Xiao-Ping Zhang

Abstract

This study seeks to advance the understanding and prediction of stock market return uncertainty through the application of advanced deep learning techniques. We introduce a novel deep learning model that utilizes a Gaussian mixture distribution to capture the complex, time-varying nature of asset return distributions in the Chinese stock market. By incorporating the Gaussian mixture distribution, our approach effectively characterizes short-term fluctuations and non-traditional features of stock returns, such as skewness and heavy tails, that are often overlooked by traditional models. Compared to GARCH models and their variants, our method demonstrates superior performance in volatility estimation, particularly during periods of heightened market volatility. It provides more accurate volatility forecasts and offers unique risk insights for different assets, thereby deepening the understanding of return uncertainty. Additionally, we propose a novel use of Code embedding which utilizes a bag-of-words approach to train hidden representations of stock codes and transforms the uncertainty attributes of stocks into high-dimensional vectors. These vectors are subsequently reduced to two dimensions, allowing the observation of similarity among different stocks. This visualization facilitates the identification of asset clusters with similar risk profiles, offering valuable insights for portfolio management and risk mitigation. Since we predict the uncertainty of returns by estimating their latent distribution, it is challenging to evaluate the return distribution when the true distribution is unobservable. However, we can measure it through the CRPS to assess how well the predicted distribution matches the true returns, and through MSE and QLIKE metrics to evaluate the error between the volatility level of the predicted distribution and proxy measures of true volatility.

Complexity vs Empirical Score

  • Math Complexity: 6.5/10
  • Empirical Rigor: 7.5/10
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper presents an innovative integration of Transformer architectures (Crossformer) with Mixture Density Networks and stock ticker embeddings to predict Chinese A-share return distributions. It features rigorous out-of-sample testing against GARCH baselines using metric standards like CRPS and QLIKE, though code availability is not explicitly specified in the excerpt.

Research Flowchart

  flowchart TD
    A["<b>Research Goal</b><br>Improve stock return uncertainty & volatility forecasting"] --> B["<b>Data & Inputs</b><br>Chinese Stock Market Returns & Stock Codes"]
    B --> C["<b>Deep Learning Framework</b><br>Fit Gaussian Mixture Distribution (GMD) to returns"]
    B --> D["<b>Code Embedding</b><br>Bag-of-words training to map stock codes to 2D risk vectors"]
    C --> E["<b>Model Evaluation</b><br>Assess fit & volatility via CRPS, MSE, and QLIKE"]
    D --> F["<b>Asset Clustering</b><br>Visualize similarities in stock risk profiles"]
    E --> G["<b>Key Findings</b><br>Outperforms GARCH models; captures heavy tails & skewness"]
    F --> G

Cited References (136)

  1. Mixture of distribution hypothesis: Analyzing daily liquidity frictions and information flows (2017)
  2. Threshold heteroskedastic models (1994)
  3. Autoregressive Conditional Heteroscedasticity with Estimates of the Variance of United Kingdom Inflation (1982)
  4. Multivariate simultaneous generalized ARCH (1995)
  5. Deep attentive survival analysis in limit order books: estimating fill probabilities with convolutional-transformers (2024)
  6. The impact of increase in minimum wages on consumer perceptions of service: A transformer model of online restaurant reviews (2021)
  7. Physics-informed convolutional transformer for predicting volatility surface (2024)
  8. Laplace-Weibull mixtures for modeling price changes (1993)
and 128 more
  1. Mixtures of Exponential Distributions (1982)
  2. Bimodal characteristic returns and predictability enhancement via machine learning (2022)
  3. A conditional gradient approach for nonparametric estimation of mixing distributions (2020)
  4. Deep learning in asset pricing (2024)
  5. Empirical asset pricing via machine learning (2020)
  6. Machine-learning-based return predictors and the spanning controversy in macro-finance (2023)
  7. Machine learning vs. economic restrictions: Evidence from stock return predictability (2023)
  8. Machine learning and portfolio optimization (2018)
  9. How Much Can Machines Learn Finance from Chinese Text Data? (2024)
  10. Modeling Asymmetric Comovements of Asset Returns (1998)
  11. Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models (2002)
  12. Information and volatility: The no-arbitrage martingale approach to timing and resolution irrelevancy (1989)
  13. Heteroskedasticity in stock returns (1990)
  14. On the relation between the expected value and the volatility of the nominal excess return on stocks (1993)
  15. Measuring and testing the impact of news on volatility (1993)
  16. Generalized autoregressive conditional heteroskedasticity (1986)
  17. Machine learning in the Chinese stock market (2022)
  18. The distribution of realized stock return volatility (2001)
  19. Modeling and Forecasting Realized Volatility (2003)
  20. Conditional Heteroskedasticity in Asset Returns: A New Approach (1991)
  21. A long memory property of stock market returns and a new model (1993)
  22. Quasi-maximum likelihood estimation and inference in dynamic models with time-varying covariances (1992)
  23. Fractionally integrated generalized autoregressive conditional heteroskedasticity (1996)
  24. A Model of Asset Trading Under the Assumption of Sequential Information Arrival (1976)
  25. Time and the Process of Security Price Adjustment (1992)
  26. Order Form and Information in Securities Markets (1991)
  27. A Subordinated Stochastic Process Model with Finite Variance for Speculative Prices (1973)
  28. Visualizing Data using t-SNE (2008)
  29. Mixtures of t-distributions for finance and forecasting (2008)
  30. Mixture density networks (1994)
  31. Short-term forecasting and uncertainty analysis of wind turbine power based on long short-term memory network and Gaussian mixture model (2019)
  32. Continuous space language models (2007)
  33. Efficient Estimation of Word Representations in Vector Space (2013)
  34. Deep embedding and alignment of protein sequences (2023)
  35. Improving distributional similarity with lessons learned from word embeddings (2015)
  36. Realized range-based estimation of integrated variance (2007)
  37. Crossformer: Transformer Utilizing Cross-Dimension Dependency for Multivariate Time Series Forecasting (2023)
  38. Additive outliers, GARCH and forecasting volatility (1999)
  39. Attention is all you need (2017)
  40. Comparing Predictive Accuracy (1995)
  41. Strictly Proper Scoring Rules, Prediction, and Estimation (2007)
  42. Forecast Accuracy Uncertainty and Momentum (2009)
  43. Bond Risk Premiums with Machine Learning (2020)
  44. Stock Return Predictability: Is it There? (2006)
  45. Risk, uncertainty, and expected returns (2016)
  46. The impact of risk and uncertainty on expected returns (2009)
  47. Unknown unknowns: Uncertainty about risk and stock returns (2018)
  48. A review of uncertainty quantification in deep learning: Techniques, applications and challenges (2021)
  49. The patent holder’s dilemma: Buy, sell, or troll? (2007) — link
  50. Deciding equivalances among conjunctive aggregate queries (2007) — link
  51. Special issue: Digital Libraries (1996)
  52. Understanding Policy-Based Networking (2001)
  53. First-Order Dynamic Logic (1979) — link
  54. The title of book one (2007) — link
  55. The title of book two (2008) — link
  56. Achieving application requirements (1990) — link
  57. Statecarts in use: structured analysis and object-orientation (1998) — link
  58. The Art of Computer Programming, Vol. 1: Fundamental Algorithms (3rd. ed.) (1997)
  59. The Art of Computer Programming (1998)
  60. An experiment in bibliographic mark-up: Parsing metadata for XML export (2010) — link
  61. Catch me, if you can: Evading network signatures with web-based polymorphic worms (2007)
  62. Catch me, if you can: Evading network signatures with web-based polymorphic worms (2008)
  63. Catch me, if you can: Evading network signatures with web-based polymorphic worms (2009)
  64. Predicate Path expressions (1979) — link
  65. LOGICS of Programs: AXIOMATICS and DESCRIPTIVE POWER (1978)
  66. Optimal Motion Control of a Ground Vehicle (2003)
  67. Algorithms for Closest-Point Problems (Computational Geometry) (1985)
  68. Introduction to Bayesian Statistics (2001) — link
  69. CLIFFORD: a Maple 11 Package for Clifford Algebra Computations, version 11 (2007) — link
  70. Stats and Analysis (2006) — link
  71. A more perfect union (2008) — link
  72. The fountain of youth (2009)
  73. Solder man (2003) — link
  74. Interview with Bill Kinder: January 13, 2005 (2005) — link
  75. The Enabling of Digital Libraries (2008)
  76. (old) Finding minimum congestion spanning trees (2000) — link
  77. (new) Finding minimum congestion spanning trees (2000) — link
  78. (old) Distributed data source verification in wireless sensor networks (2009) — link
  79. (new) Distributed data source verification in wireless sensor networks (2009) — link
  80. Portalis: using competitive online interactions to support aid initiatives for the homeless (2008) — link
  81. Visual Basic 6: Design, Specification, and Objects with Other (1999)
  82. The Latex Web Companion: Integrating TEX, HTML, and XML (1999)
  83. Vertex Types in Book-Embeddings (1987)
  84. Vertex Types in Book-Embeddings (1987)
  85. CHI ‘08: CHI ‘08 extended abstracts on Human factors in computing systems (2008)
  86. Algorithms for Closest-Point Problems (Computational Geometry) (1985)
  87. IEEE TCSC Executive Committee (2004) — link
  88. Distributed systems (2nd Ed.) (1993)
  89. New Algorithms for Dependency-Directed Backtracking (Master’s thesis) (1986)
  90. New Algorithms for Dependency-Directed Backtracking (Master’s thesis) (1986)
  91. Seminumerical Algorithms (1981)
  92. E-commerce and cultural values (2001) — link
  93. Chapter 9 (2002) — link
  94. The implementation of electronic commerce in SMEs in Singapore (Incoll) (2003) — link
  95. E-commerce and cultural values - (InBook-num-in-chap) (2004) — link
  96. E-commerce and cultural values (Inbook-text-in-chap) (2005) — link
  97. E-commerce and cultural values (Inbook-num chap) (2006) — link
  98. A library-based synthesis methodology for reversible logic (2010)
  99. Synthesis of Reversible Circuit Using Cycle-Based Approach (2010)
  100. Algorithmic Enumeration of Ideal Classes for Quaternion Orders (2010) — link
  101. Chapter II: Notes on data structuring (1972) — link
  102. Transcript of question and answer session (1981) — link
  103. Go to statement considered harmful (1979) — link
  104. Three-dimensional virtual acoustic displays (1992) — link
  105. Managerial expert systems and organizational change: some critical research issues (1987) — link
  106. Simplified Structured COBOL with Microsoft/MicroFocus COBOL (1990)
  107. A multi-radio unification protocol for IEEE 802.11 wireless networks (2004)
  108. Wireless Sensor Networks: A Survey (2002)
  109. A Survey on Wireless Multimedia Sensor Networks (2007)
  110. SSCH : Slotted Seeded Channel Hopping for Capacity Improvement in IEEE 802.11 Ad-Hoc Wireless Networks (2004)
  111. XBOW Sensor Motes Specifications (2008)
  112. Overview of Sensor Networks (2004)
  113. CodeBlue : Sensor Networks for Medical Care (2008)
  114. Investigating Network Architectures for Body Sensor Networks (2007) — link
  115. Channel-Hopping Multiple Access (2000)
  116. Body Sensor Networks (2008)
  117. Mapping Powerlists onto Hypercubes (1994)
  118. Automatic Parallelization for Distributed-Memory Multiprocessing Systems (1989)
  119. User recovery and reversal in interactive systems (1984)
  120. Generalizing specifications for uniformly implemented loops (1985)
  121. Towards monolingual programming environments (1985)
  122. The \TeX book (1984)
  123. Distributed algorithms for finding centers and medians in networks (1984)
  124. \it \LaTeX : A Document Preparation System (1986)
  125. Program transformations in a denotational setting (1985)
  126. Seminumerical Algorithms (1981)
  127. A high-level approach to computer document formatting (1980)
  128. A multifrequency MAC specially designed for wireless sensor network applications (2010) — link

0 of 136 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.