References and further reading

These references explain the statistical ideas behind the guides. A citation is not a claim that Trex implements every estimator, assumption, inference procedure or computational refinement in the cited work. The API reference and each guide’s scope section describe the implementation actually provided.

Regression and moment restrictions

  • Hansen, L. P. (1982). Large Sample Properties of Generalized Method of Moments Estimators. Econometrica 50(4), 1029–1054. DOI.
  • Newey, W. K., and K. D. West (1987). A Simple, Positive Semi-Definite, Heteroskedasticity and Autocorrelation Consistent Covariance Matrix. Econometrica 55(3), 703–708. DOI.
  • Newey, W. K., and R. J. Smith (2004). Higher Order Properties of GMM and Generalized Empirical Likelihood Estimators. Econometrica 72(1), 219–255. DOI.
  • Muandet, K., W. Jitkrittum, and J. Kübler (2020). Kernel Conditional Moment Test via Maximum Moment Restriction. UAI. Proceedings.
  • Optimal Convex M-Estimation for Linear Regression via Antitonic Score Matching. The R package asm documents the original procedure and references. Trex ports its core score-learning idea; its numerical options and inferential implementation are not a claim of complete feature parity with the R package.

For least squares, Frisch–Waugh–Lovell projections, sandwich covariance and identification, a graduate econometrics treatment is useful alongside the worked code. GMM and kernel criteria require valid moments; automatic differentiation supplies derivatives, not identifying assumptions.

Choice and structural dynamics

  • Rust, J. (1987). Optimal Replacement of GMC Bus Engines: An Empirical Model of Harold Zurcher. Econometrica 55(5), 999–1033. DOI.
  • Hotz, V. J., and R. A. Miller (1993). Conditional Choice Probabilities and the Estimation of Dynamic Models. Review of Economic Studies 60(3), 497–529. DOI.
  • Train, K. (2009). Discrete Choice Methods with Simulation, second edition. Author’s book page.
  • Kallus, N., and M. Udell (2016). Revealed Preference at Scale: Learning Personalized Preferences from Assortment Choices. arXiv.

Panels and low-rank structure

  • Arkhangelsky, D., S. Athey, D. A. Hirshberg, G. W. Imbens, and S. Wager (2021). Synthetic Difference-in-Differences. American Economic Review 111(12), 4088–4118. DOI. Reference implementation.
  • Athey, S., M. Bayati, N. Doudchenko, G. Imbens, and K. Khosravi (2021). Matrix Completion Methods for Causal Panel Data Models. JASA. Working-paper version.
  • Mazumder, R., T. Hastie, and R. Tibshirani (2010). Spectral Regularization Algorithms for Learning Large Incomplete Matrices. JMLR 11, 2287–2322. Article.
  • Bonhomme, S., T. Lamadon, and E. Manresa (2022). Discretizing Unobserved Heterogeneity. Econometrica 90(2), 625–643. Working-paper version.

Generative models

  • Gulrajani, I., F. Ahmed, M. Arjovsky, V. Dumoulin, and A. Courville (2017). Improved Training of Wasserstein GANs. arXiv.
  • Ho, J., A. Jain, and P. Abbeel (2020). Denoising Diffusion Probabilistic Models. arXiv.
  • Sohn and Song. Parallelly Tempered Generative Adversarial Nets: Toward Stabilized Gradients. See the paper and current revision for the convex tempering and coherency construction used by TabularPTGAN.
  • Hu, E. J., et al. (2022). LoRA: Low-Rank Adaptation of Large Language Models. arXiv.
  • Dettmers, T., A. Pagnoni, A. Holtzman, and L. Zettlemoyer (2023). QLoRA: Efficient Finetuning of Quantized LLMs. arXiv.

Numerical references used by the repository

The test suite compares selected calculations with SciPy, Statsmodels, linearmodels, saved R fixtures, independent finite differences and the synthdid California example. See tests and fixture provenance. These checks have defined scopes; they do not establish correctness of all possible inputs or replace an estimator-specific design and validation study.