Paper: arXiv 2312.17375

Abstract

This paper introduces a new causal structure learning method for nonstationary time series data, a common data type found in fields such as finance, economics, healthcare, and environmental science. Our work builds upon the constraint-based causal discovery from nonstationary data algorithm (CD-NOD). We introduce a refined version (CD-NOTS) which is designed specifically to account for lagged dependencies in time series data. We compare the performance of different algorithmic choices, such as the type of conditional independence test and the significance level, to help select the best hyperparameters given various scenarios of sample size, problem dimensionality, and availability of computational resources. Using the results from the simulated data, we apply CD-NOTS to a broad range of real-world financial applications in order to identify causal connections among nonstationary time series data, thereby illustrating applications in factor-based investing, portfolio diversification, and comprehension of market dynamics.

Complexity vs Empirical Score

  • Math Complexity: 8.0/10
  • Empirical Rigor: 6.0/10 (capped from 7.0: transaction costs not modelled; no out-of-sample test; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper presents advanced mathematical concepts such as conditional independence tests (KCIT, RCoT, CMIknn) and nonparametric causal discovery algorithms, requiring high mathematical density. It demonstrates strong empirical rigor by evaluating the algorithm on simulated and real-world financial data, addressing implementation details like hyperparameter selection and computational resource considerations.

Research Flowchart

  flowchart TD
  A["Research Goal: <br>Discover causal structures in <br>nonstationary financial time-series data"] --> B["Data Input:<br>Simulated & Real Financial Data"]
  B --> C["Core Methodology:<br>CD-NOTS Algorithm<br>(Lagged Dependencies)"]
  C --> D["Hyperparameter Tuning:<br>Conditional Independence Tests & Significance"]
  D --> E["Validation:<br>Performance on Simulated Data"]
  E --> F["Key Findings & Outcomes:"]
  F --> G["Factor-based Investing<br>Portfolio Diversification<br>Market Dynamics"]

Cited References (73)

  1. Approximating discrete probability distributions with dependence trees (1968)
  2. Probabilistic R easoning in I ntelligent S ystems: N etworks of P lausible I nference (1988)
  3. Causal Discovery from Heterogeneous/Nonstationary Data with Independent Changes (2020) — link
  4. Kernel-based Conditional Independence Test and Application in Causal Discovery (2012) — link
  5. Multi-domain Causal Structure Learning in Linear Systems (2018) — link
  6. Causality: Models , Reasoning, and Inference (2000)
  7. A Kernel Statistical Test of Independence (2007) — link
  8. Causation, Prediction, and Search (2001) — link
and 65 more
  1. Investigating Causal Relations by Econometric Models and Cross-spectral Methods (1969)
  2. A Comparison of Public Causal Search Packages on Linear, Gaussian Data with No Latent Variables (2017)
  3. On the Identifiability of the Post-Nonlinear Causal Model (2009)
  4. Behind Distribution Shift: Mining Driving Forces of Changes and Causal Arrows (2017) — link
  5. Causal Discovery from Nonstationary/Heterogeneous Data: Skeleton Estimation and Orientation Determination (2017) — link
  6. Causal inference using invariant prediction: identification and confidence intervals (2015) — link
  7. Estimating causal effects of treatments in randomized and nonrandomized studies (1974) — link
  8. Causality (2009)
  9. The Deductive Approach to Causal Inference (2014) — link
  10. The Causal-Neural Connection: Expressiveness, Learnability, and Inference (2021) — link
  11. On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9 (1990) — link
  12. A Survey on Causal Inference (2021) — link
  13. Constructing Bayesian Network Models of Gene Expression Networks from Microarray Data (2000) — link
  14. Variable selection in high-dimensional linear models: partially faithful distributions and the PC-simple algorithm (2009) — link
  15. Adjacency-Faithfulness and Conservative Causal Inference (2012) — link
  16. The Max-Min Hill-Climbing Bayesian Network Structure Learning Algorithm (2006) — link
  17. Iterative Causal Discovery in the Possible Presence of Latent Confounders and Selection Bias (2021) — link
  18. Learning high-dimensional directed acyclic graphs with latent and selection variables (2012) — link
  19. A Survey of Methods, Challenges and Perspectives in Causality (2023) — link
  20. A Linear Non-Gaussian Acyclic Model for Causal Discovery (2006) — link
  21. On the Identifiability of the Post-Nonlinear Causal Model (2012) — link
  22. Distinguishing Causes from Effects using Nonlinear Acyclic Causal Models (2010) — link
  23. From Statistical to Causal Learning (2022) — link
  24. Causal Inference for Time series Analysis: Problems, Methods and Evaluation (2021) — link
  25. Investigating Causal Relations by Econometric Models and Cross-Spectral Methods (2001)
  26. Methods and tools for causal discovery and causal inference (2022) — link
  27. A mathematical theory of communication (1948) — link
  28. Synchronization as adjustment of information rates: Detection from bivariate time series (2001) — link
  29. Measuring Information Transfer (2000) — link
  30. On Causal Discovery from Time Series Data using FCI (2010)
  31. Detecting and quantifying causal associations in large nonlinear time series datasets (2019) — link
  32. A permutation-based kernel conditional independence test (2014)
  33. Fast Conditional Independence Test for Vector Variables with Large Sample Sizes (2018) — link
  34. Kernel-based Tests for Joint Independence (2016) — link
  35. Towards Causal Representation Learning (2021) — link
  36. Causal inference using invariant prediction: identification and confidence intervals (2015) — link
  37. A Kernel Statistical Test of Independence (2007) — link
  38. Approximate Kernel-based Conditional Independence Tests for Fast Non-Parametric Causal Discovery (2017) — link
  39. Conditional independence testing based on a nearest-neighbor estimator of conditional mutual information (2017) — link
  40. Inferring the directionality of coupling with conditional mutual information (2008) — link
  41. Causal Discovery with Continuous Additive Noise Models (2014) — link
  42. An Approximate Distribution of Estimates of Variance Components (1946)
  43. The Significance of the Difference Between Two Means when the Population Variances are Unequal (1938)
  44. Chi Squared Approximations to the Distribution of a Sum of Independent Random Variables (1983) — link
  45. AN APPROXIMATION TO THE DISTRIBUTION OF QUADRATIC FORMS IN NORMAL RANDOM VARIABLES (1988) — link
  46. A comparison of efficient approximations for a weighted sum of chi-squared random variables (2015) — link
  47. PARTIAL CORRELATION AND CONDITIONAL CORRELATION AS MEASURES OF CONDITIONAL INDEPENDENCE (2004) — link
  48. A simple approach for finding the globally optimal Bayesian network structure (2012)
  49. Optimal Structure Identification with Greedy Search (2003) — link
  50. Greedy relaxations of the sparsest permutation algorithm (2022) — link
  51. Estimation of a Structural Vector Autoregression Model Using Non-Gaussianity (2010) — link
  52. DirectLiNGAM: A Direct Method for Learning a Linear Non-Gaussian Structural Equation Model (2011) — link
  53. Estimating the Dimension of a Model (1978) — link
  54. Theory Refinement on Bayesian Networks (1991)
  55. Common risk factors in the returns on stocks and bonds (1993) — link
  56. Causal Inference and Causal Explanation with Background Knowledge (1995)
  57. Growth vs value investing: Persistence and time trend before and after COVID-19 (2023) — link
  58. Causal Factor Investing: Can Factor Investing Become Scientific? (2023)
  59. Causality: Models, Reasoning and Inference (2009)
  60. Constraint-based causal discovery with mixed data (2018) — link
  61. Robust Causal Structure Learning with Some Hidden Variables (2019) — link
  62. A new method for estimating the probability of causal relationships from observational data: Application to the study of the short-term effects of air pollution on cardiovascular and respiratory disease (2023) — link
  63. Causal Discovery with Unobserved Confounding and Non-Gaussian Data (2023) — link
  64. Discrepancy-Based Theory and Algorithms for Forecasting Non-Stationary Time Series (2020) — link
  65. A weaker faithfulness assumption based on triple interactions (2021) — link

0 of 73 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.