Paper: arXiv 2412.06417

Abstract

Financial time series (FTS) generation models are a core pillar to applications in finance. Risk management and portfolio optimization rely on realistic multivariate price generation models. Accordingly, there is a strong modelling literature dating back to Bachelier’s Theory of Speculation in 1901. Generating FTS using deep generative models (DGMs) is still in its infancy. In this work, we systematically compare DGMs against state-of-the-art parametric alternatives for multivariate FTS generation. We initially compare both DGMs and parametric models over increasingly complex synthetic datasets. The models are evaluated through distance measures for varying distribution moments of both the full and rolling FTS. We then apply the best performing DGM models to empirical data, demonstrating the benefit of DGMs through a implied volatility trading task.

Complexity vs Empirical Score

  • Math Complexity: 7.5/10
  • Empirical Rigor: 6.0/10 (capped from 7.0: transaction costs not modelled; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper presents advanced mathematical concepts such as multivariate GARCH decompositions, factor stochastic volatility models, and various deep generative model architectures, indicating high math complexity. It also demonstrates strong empirical rigor through systematic comparisons on synthetic and empirical datasets, evaluating multiple models and applying the best to a specific trading task (implied volatility trading), which shows backtest-ready and implementation-heavy work.

Research Flowchart

  flowchart TD
  A["Research Goal: Systematically Compare DGMs vs Parametric Models for Multivariate FTS Generation"] --> B["Methodology: 2-Phase Experiment"]
  B --> C["Phase 1: Synthetic Data"]
  C --> D["Computational Process: Evaluate DGMs & Parametric Models<br>via Distance Measures on Distribution Moments"]
  B --> E["Phase 2: Empirical Data"]
  E --> D
  D --> F["Outcomes: DGMs Surpass Parametric Models"]
  F --> G["Applied Outcome: Successful Implied Volatility Trading"]

Cited References (209)

  1. Tackling the Exponential Scaling of Signature-Based Generative Adversarial Networks for High-Dimensional Financial Time-Series Generation (2022)
  2. Conditional Sig-Wasserstein GANs for Time Series Generation (2020)
  3. Time-series Generative Adversarial Networks (2019)
  4. Universal features of price formation in financial markets: perspectives from deep learning (2021)
  5. Generative moment matching networks (2015)
  6. Efficient Bayesian inference for multivariate factor stochastic volatility models (2017)
  7. Network momentum across asset classes — → 2308.11294
  8. Density estimation using Real NVP (2022)
and 201 more
  1. rmgarch: Multivariate GARCH models. (2022)
  2. Modelling the coherence in short-run nominal exchange rates: a multivariate generalized ARCH model (1990)
  3. A multivariate generalized autoregressive conditional heteroscedasticity model with time-varying correlations (2002)
  4. Dynamic conditional correlation: A simple class of multivariate generalized autoregressive conditional heteroskedasticity models (2002)
  5. The copula-garch model of conditional dependencies: An international stock market application (2006)
  6. Multivariate GARCH models (2009)
  7. Multivariate GARCH models: a survey (2006)
  8. Network Filtering of Spatial-temporal GNN for Multivariate Time-series Prediction (2022)
  9. Multivariate realized volatility forecasting with graph neural network (2022)
  10. A practical guide to harnessing the HAR volatility model (2021)
  11. Generation of Realistic Synthetic Financial Time-series (2022)
  12. Neural networks generative models for time series (2022)
  13. Synthetic data generation with deep generative models to enhance predictive tasks in trading strategies (2022)
  14. Synthetic Data Augmentation for Deep Reinforcement Learning in Financial Trading (2022)
  15. Generative Machine Learning for Multivariate Equity Returns — → 2311.14735
  16. On Correlated Stock Market Time Series Generation (2023)
  17. Quant GANs: deep generation of financial time series (2019)
  18. Deep Generative Modeling for Financial Time Series with Application in VaR: A Comparative Review (2024)
  19. FactorVAE: A Probabilistic Dynamic Factor Model Based on Variational Autoencoder for Predicting Cross-Sectional Stock Returns (2022)
  20. Auto-encoding variational Bayes (2014)
  21. Forecasting Dynamic Term Structure Models with Autoencoders (2021)
  22. Learning long-term dependencies in NARX recurrent neural networks (1996)
  23. Distilling Financial Models by Symbolic Regression (2021)
  24. PCF-GAN: generating sequential data via the characteristic function of measures on the path space (2023) — link
  25. Conditional image generation with pixelcnn decoders (2016)
  26. Wavenet: A generative model for raw audio (2016) — link
  27. Generative adversarial networks (2020)
  28. Variational inference with normalizing flows (2015)
  29. Neural Copula: A unified framework for estimating generic high-dimensional Copula functions (2022) — link
  30. Unsupervised Representation Learning for Time Series: A Review (2023) — link
  31. Decision-aware conditional gans for time series data (2020) — link
  32. Conditional GAN for timeseries generation (2020) — link
  33. Learning who is in the market from time series: market participant discovery through adversarial calibration of multi-agent simulators (2021) — link
  34. MTCopula: Synthetic complex data generation using copula (2021)
  35. Generating synthetic data in finance: opportunities, challenges and pitfalls (2020)
  36. Tail-gan: Nonparametric scenario generation for tail risk estimation (2022) — link
  37. Score-Based Generative Modeling through Stochastic Differential Equations
  38. Pagan: Portfolio analysis with generative adversarial networks (2019) — link
  39. Generation of time series and reinforcement learning (2022)
  40. Autoregressive denoising diffusion models for multivariate probabilistic time series forecasting (2021)
  41. Scenario generation for market risk models using generative neural networks (2022)
  42. Tackling the Exponential Scaling of Signature-Based Generative Adversarial Networks for High-Dimensional Financial Time-Series Generation (2022)
  43. Financial Time Series Data Augmentation with Generative Adversarial Networks and extended Intertemporal Return Plots (2022) — link
  44. Quantile regression (2001)
  45. Dropout as a bayesian approximation: Representing model uncertainty in deep learning (2016)
  46. Bayesian neural networks (1997)
  47. Hierarchical modelling for financial data (2020)
  48. Generation of Synthetic Financial Time-Series with Generative Adversarial Networks (2020)
  49. Financial forecasting with Neural Networks and Reservoir Computing
  50. Generative Models for Financial Data (2019)
  51. Cholesky–ANN models for predicting multivariate realized volatility (2019)
  52. Realized Volatility Forecasting with Neural Networks (2020)
  53. Deep learning and time series-to-image encoding for financial forecasting (2020)
  54. Deep Bayesian Estimation of the Rough Path-Dependent Volatility Model for Forecasting Purposes (2023)
  55. World Models (2018)
  56. Professor Forcing: A New Algorithm for Training Recurrent Networks (2016)
  57. VFunc: a Deep Generative Model for Functions (2018)
  58. PCF-GAN: generating sequential data via the characteristic function of measures on the path space (2023)
  59. Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks (2020)
  60. The Neural Hawkes Process: A Neurally Self-Modulating Multivariate Point Process (2016)
  61. Universal features of price formation in financial markets: perspectives from deep learning (2018)
  62. Time-series Generative Adversarial Networks (2019)
  63. Modeling financial time-series with generative adversarial networks (2019)
  64. Generative adversarial networks in time series: A systematic literature review (2023)
  65. The market generator (2019)
  66. Adversarial generalized method of moments (2018) — link
  67. Nonlinear principal component analysis using autoassociative neural networks (1991)
  68. Discovering governing equations from data by sparse identification of nonlinear dynamical systems (2015)
  69. TimeVAE: A Variational Auto-Encoder for Multivariate Time Series Generation (2021)
  70. Learning Stochastic Dynamics with Statistics-Informed Neural Network (2022)
  71. Pareto GAN: Extending the Representational Power of GANs to Heavy-Tailed Distributions (2021)
  72. GANStick: US stock forecasting with GAN-generated candlesticks (2020)
  73. Trading via image classification (2020)
  74. From time series to complex networks: The visibility graph (2008)
  75. Sparse identification of nonlinear dynamics for model predictive control in the low-data limit (2018)
  76. Mixture density networks (1994)
  77. Volatility prediction with mixture density networks (1998)
  78. Temporal convolutional networks for action segmentation and detection (2017)
  79. Neural autoregressive distribution estimation (2016)
  80. Learning Curve Dynamics with Artificial Neural Networks (2018)
  81. Model Calibration with Neural Networks (2016)
  82. Calibrating rough volatility models: a convolutional neural network approach (2018)
  83. Deep calibration of rough stochastic volatility models (2018)
  84. Volatility Model Calibration With Convolutional Neural Networks (2018)
  85. Displacement interpolation using Lagrangian mass transport (2011)
  86. Topological Portfolio Selection and Optimization — → 2310.14881
  87. Towards Controllable Time Series Generation (2024) — link
  88. A simple approximate long-memory model of realized volatility (2009)
  89. Volatility trading (2013)
  90. Deep learning volatility: a deep neural network perspective on pricing and calibration in (rough) volatility models (2019)
  91. Sig-Wasserstein GANs for time series generation (2021)
  92. A parsimonious multi-asset Heston model: Calibration and derivative pricing (2011)
  93. Long short-term memory (1997)
  94. Low-dose CT image denoising using a generative adversarial network with Wasserstein distance and perceptual loss (2018)
  95. Simulation of arbitrage-free implied volatility surfaces (2022)
  96. Copula flows for synthetic data generation (2021) — link
  97. Autoregressive quantile networks for generative modeling (2018)
  98. A distributional perspective on reinforcement learning (2017)
  99. DeepAR: Probabilistic forecasting with autoregressive recurrent networks (2020)
  100. cCorrGAN: Conditional correlation GAN for learning empirical conditional distributions in the elliptope (2021)
  101. MAS-GAN: Adversarial Calibration of Multi-Agent Market Simulators. (2020)
  102. Corrgan: Sampling realistic financial correlation matrices using generative adversarial networks (2020)
  103. Quantnet: Transferring learning across systematic trading strategies (2020) — link
  104. Time-series generation by contrastive imitation (2021)
  105. Cot-gan: Generating sequential data via causal optimal transport (2020)
  106. Empirical properties of asset returns: stylized facts and statistical issues (2001)
  107. Probabilistic forecasting of sensory data with generative adversarial networks–forgan (2019)
  108. Fin-gan: Forecasting and classifying financial time series via generative adversarial networks (2024)
  109. Path Shadowing Monte-Carlo — → 2308.01486
  110. Wasserstein generative adversarial networks (2017)
  111. Quant GANs: deep generation of financial time series (2020)
  112. Real-valued (medical) time series generation with recurrent conditional gans (2017) — link
  113. Score-based generative modeling through stochastic differential equations (2020) — link
  114. Rough paths, signatures and the modelling of functions on streams (2014) — link
  115. An experiment in bibliographic mark-up: Parsing metadata for XML export (2010) — link
  116. Catch me, if you can: Evading network signatures with web-based polymorphic worms (2007)
  117. Catch me, if you can: Evading network signatures with web-based polymorphic worms (2008)
  118. Catch me, if you can: Evading network signatures with web-based polymorphic worms (2009)
  119. Predicate Path expressions (1979) — link
  120. LOGICS of Programs: AXIOMATICS and DESCRIPTIVE POWER (1978)
  121. Optimal Motion Control of a Ground Vehicle (2003)
  122. Algorithms for Closest-Point Problems (Computational Geometry) (1985)
  123. Introduction to Bayesian Statistics (2001) — link
  124. CLIFFORD: a Maple 11 Package for Clifford Algebra Computations, version 11 (2007) — link
  125. Stats and Analysis (2006) — link
  126. A more perfect union (2008) — link
  127. The fountain of youth (2009)
  128. Solder man (2003) — link
  129. Interview with Bill Kinder: January 13, 2005 (2005) — link
  130. The Enabling of Digital Libraries (2008)
  131. (old) Finding minimum congestion spanning trees (2000) — link
  132. (new) Finding minimum congestion spanning trees (2000) — link
  133. (old) Distributed data source verification in wireless sensor networks (2009) — link
  134. (new) Distributed data source verification in wireless sensor networks (2009) — link
  135. Portalis: using competitive online interactions to support aid initiatives for the homeless (2008) — link
  136. Visual Basic 6: Design, Specification, and Objects with Other (1999)
  137. The Latex Web Companion: Integrating TEX, HTML, and XML (1999)
  138. Vertex Types in Book-Embeddings (1987)
  139. Vertex Types in Book-Embeddings (1987)
  140. CHI ‘08: CHI ‘08 extended abstracts on Human factors in computing systems (2008)
  141. Algorithms for Closest-Point Problems (Computational Geometry) (1985)
  142. IEEE TCSC Executive Committee (2004) — link
  143. Distributed systems (2nd Ed.) (1993)
  144. New Algorithms for Dependency-Directed Backtracking (Master’s thesis) (1986)
  145. New Algorithms for Dependency-Directed Backtracking (Master’s thesis) (1986)
  146. Seminumerical Algorithms (1981)
  147. E-commerce and cultural values (2001) — link
  148. Chapter 9 (2002) — link
  149. The implementation of electronic commerce in SMEs in Singapore (Incoll) (2003) — link
  150. E-commerce and cultural values - (InBook-num-in-chap) (2004) — link
  151. E-commerce and cultural values (Inbook-text-in-chap) (2005) — link
  152. E-commerce and cultural values (Inbook-num chap) (2006) — link
  153. A library-based synthesis methodology for reversible logic (2010)
  154. Synthesis of Reversible Circuit Using Cycle-Based Approach (2010)
  155. Algorithmic Enumeration of Ideal Classes for Quaternion Orders (2010) — link
  156. Chapter II: Notes on data structuring (1972) — link
  157. Transcript of question and answer session (1981) — link
  158. Go to statement considered harmful (1979) — link
  159. Three-dimensional virtual acoustic displays (1992) — link
  160. Managerial expert systems and organizational change: some critical research issues (1987) — link
  161. Simplified Structured COBOL with Microsoft/MicroFocus COBOL (1990)
  162. A multi-radio unification protocol for IEEE 802.11 wireless networks (2004)
  163. Wireless Sensor Networks: A Survey (2002)
  164. A Survey on Wireless Multimedia Sensor Networks (2007)
  165. SSCH : Slotted Seeded Channel Hopping for Capacity Improvement in IEEE 802.11 Ad-Hoc Wireless Networks (2004)
  166. XBOW Sensor Motes Specifications (2008)
  167. Overview of Sensor Networks (2004)
  168. CodeBlue : Sensor Networks for Medical Care (2008)
  169. Investigating Network Architectures for Body Sensor Networks (2007) — link
  170. Channel-Hopping Multiple Access (2000)
  171. Body Sensor Networks (2008)
  172. Mapping Powerlists onto Hypercubes (1994)
  173. Automatic Parallelization for Distributed-Memory Multiprocessing Systems (1989)
  174. User recovery and reversal in interactive systems (1984)
  175. Generalizing specifications for uniformly implemented loops (1985)
  176. Towards monolingual programming environments (1985)
  177. The \TeX book (1984)
  178. Distributed algorithms for finding centers and medians in networks (1984)
  179. \it \LaTeX : A Document Preparation System (1986)
  180. Program transformations in a denotational setting (1985)
  181. Seminumerical Algorithms (1981)
  182. Th 'e orie math 'e matique du jeu (1901)
  183. A high-level approach to computer document formatting (1980)
  184. A multifrequency MAC specially designed for wireless sensor network applications (2010) — link
  185. Modeling Univariate and Multivariate Stochastic Volatility in R with stochvol and factorstochvol (2021) — link
  186. Incorporating Corporation Relationship via Graph Convolutional Neural Networks for Stock Price Prediction (2018)
  187. Financial time series forecasting with multi-modality graph neural network (2022)
  188. Temporal and Heterogeneous Graph Neural Network for Financial Time Series Prediction — → 2305.08740
  189. Institutional members of the \TeX Users Group (2017) — link
  190. acmart— Class for typesetting publications of ACM (2017) — link
  191. The \textsl doclicense package (2022) — link
  192. Reasoning About Naming Systems (1993) — link
  193. Babel, a Multilingual Style-Option System for Use with LaTeX’s Standard Document Styles (1991)
  194. A Methodology for Implementing Highly Concurrent Data Objects (1993) — link
  195. Publication quality tables in \LaTeX (2005)
  196. Using the amsthm Package (2015)
  197. R: A Language and Environment for Statistical Computing (2019) — link
  198. UMass Citation Field Extraction Dataset (2013) — link
  199. Citation concept analysis (CCA) —A new form of citation analysis revealing the usefulness of concepts for other researchers illustrated by two exemplary case studies including classic books by Thomas S.~Kuhn and Karl R.~Popper (2019) — link
  200. Learning Soft Linear Constraints with Application to Citation Field Extraction (2014) — link
  201. Maintaining Discrete Probability Distributions Optimally (1993)

5 of 209 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.