Drift-Adaptive Non-Crossing Quantile Regression With Local Conformal Calibration For Nonstationary Data
Keywords:
Conformal calibration, Distributional drift, Non-crossing quantiles, Quantile regression, Prediction intervals.Abstract
This article proposes Drift-Adaptive Non-Crossing Quantile Regression (DAN-CQR), a methodological framework for modeling conditional distributions under nonstationary and tail-sensitive data. Conventional quantile regression can describe heterogeneous effects across the response distribution, but it often treats all observations as equally relevant, may produce crossing quantile curves, and usually relies on global conformal corrections that are less efficient under distributional drift and local heteroscedasticity. DAN-CQR integrates four components: memory-weighted composite quantile loss, residual-adaptive robustness, non-crossing rearrangement, and local conformal calibration. A simulation study with nonlinear structure, heteroscedasticity, heavy-tailed asymmetric errors, outliers, and regime changes was conducted to assess its behavior. The preliminary results show that DAN-CQR achieves calibrated coverage of 0.950 with a narrower average interval width of 11.460 and lower median absolute error than the global conformalized linear and polynomial quantile regression baselines. These findings suggest that the proposed framework can provide coherent quantile estimates and adaptive prediction bands for dynamic data. The method offers a promising direction for robust and interpretable distributional regression in economics, finance, environmental risk, health, education, and public policy.
References
[1] Koenker and G. Bassett, Jr., “Regression quantiles,” Econometrica, vol. 46, no. 1, pp. 33–50, 1978, doi: 10.2307/1913643.
[2] Koenker and K. F. Hallock, “Quantile regression,” Journal of Economic Perspectives, vol. 15, no. 4, pp. 143–156, 2001, doi: 10.1257/jep.15.4.143.
[3] Koenker, Quantile Regression. Cambridge, U.K.: Cambridge University Press, 2005, doi: 10.1017/CBO9780511754098.
[4] Koenker, “Quantile regression: 40 years on,” Annual Review of Economics, vol. 9, pp. 155–176, 2017, doi: 10.1146/annurev economics-063016-103651.
[5] K. Yu and R. A. Moyeed, “Bayesian quantile regression,” Statistics & Probability Letters, vol. 54, no. 4, pp. 437–447, 2001, doi: 10.1016/S0167-7152(01)00124-9.
[6] Koenker, “Quantile regression for longitudinal data,” Journal of Multivariate Analysis, vol. 91, no. 1, pp. 74–89, 2004, doi:10.1016/j.jmva.2004.05.006.
[7] Portnoy, “Censored regression quantiles,” Journal of the American Statistical Association, vol. 98, no. 464, pp. 1001–1012, 2003, doi: 10.1198/016214503000000954.
[8] Meinshausen, “Quantile regression forests,” Journal of Machine Learning Research, vol. 7, pp. 983–999, 2006.
[9] Belloni and V. Chernozhukov, “L1-penalized quantile regression in high-dimensional sparse models,” The Annals of Statistics, vol. 39, no. 1, pp. 82–130, 2011, doi: 10.1214/10-AOS827.
[10] Zou and M. Yuan, “Composite quantile regression and the oracle model selection theory,” The Annals of Statistics, vol. 36, no. 3, pp. 1108–1126, 2008, doi: 10.1214/07-AOS507.
[11] Chernozhukov, I. Fernández-Val, and A. Galichon, “Quantile and probability curves without crossing,” Econometrica, vol. 78, no. 3, pp. 1093–1125, 2010, doi: 10.3982/ECTA7880.
[12] D. Bondell, B. J. Reich, and H. Wang, “Noncrossing quantile regression curve estimation,” Biometrika, vol. 97, no. 4, pp. 825–838, 2010, doi: 10.1093/biomet/asq048.
[13] M. Fasiolo, S. N. Wood, M. Zaffran, R. Nedellec, and Y. Goude, “qgam: Bayesian nonparametric quantile regression modeling in R,” Journal of Statistical Software, vol. 100, no. 9, pp. 1–31, 2021, doi: 10.18637/jss.v100.i09.
[14] X. He, X. Pan, K. M. Tan, and W.-X. Zhou, “Smoothed quantile regression with large-scale inference,” Journal of Econometrics, vol. 232, no. 2, pp. 367–388, 2023, doi: 10.1016/j.jeconom.2021.07.010.
[15] K. M. Tan, L. Wang, and W.-X. Zhou, “High-dimensional quantile regression: Convolution smoothing and concave regularization,” Journal of the Royal Statistical Society: Series B, vol. 84, no. 1, pp. 205–233, 2022, doi: 10.1111/rssb.12485.
[16] Y. Romano, E. Patterson, and E. J. Candès, “Conformalized quantile regression,” in Advances in Neural Information Processing Systems, vol. 32, 2019, pp. 3543–3553.
[17] J. Lei, M. G’Sell, A. Rinaldo, R. J. Tibshirani, and L. Wasserman, “Distribution-free predictive inference for regression,” Journal of the American Statistical Association, vol. 113, no. 523, pp. 1094–1111, 2018, doi: 10.1080/01621459.2017.1307116.
[18] A. Brando, J. Gimeno, J. A. Rodríguez-Serrano, and J. Vitrià, “Deep non-crossing quantiles through the partial derivative,” in Proceedings of the 25th International Conference on Artificial Intelligence and Statistics, PMLR, vol. 151, 2022, pp. 7902–7912.
[19] A. J. Cannon, “Non-crossing nonlinear regression quantiles by monotone composite quantile regression neural network, with application to rainfall extremes,” Stochastic Environmental Research and Risk Assessment, vol. 32, pp. 3207–3225, 2018, doi: 10.1007/s00477-018-1573-6.
[20] G. Carlier, V. Chernozhukov, and A. Galichon, “Vector quantile regression: An optimal transport approach,” The Annals of Statistics, vol. 44, no. 3, pp. 1165–1192, 2016, doi: 10.1214/15-AOS1401.
[21] G. Carlier, V. Chernozhukov, and A. Galichon, “Vector quantile regression beyond the specified case,” Journal of Multivariate Analysis, vol. 161, pp. 96–102, 2017, doi: 10.1016/j.jmva.2017.07.001.
[22] R. Koenker and Z. Xiao, “Quantile autoregression,” Journal of the American Statistical Association, vol. 101, no. 475, pp. 980–990, 2006, doi: 10.1198/016214506000000672.
[23] I. Takeuchi, Q. V. Le, T. D. Sears, and A. J. Smola, “Nonparametric quantile estimation,” Journal of Machine Learning Research, vol. 7, pp. 1231–1264, 2006.
[24] Y. Wu and Y. Liu, “Stepwise multiple quantile regression estimation using non-crossing constraints,” Statistics and Its Interface, vol. 2, no. 3, pp. 299–310, 2009.
[25] Chen, W. Liu, and Y. Zhang, “Quantile regression under memory constraint,” The Annals of Statistics, vol. 47, no. 6, pp. 3244–3273, 2019, doi: 10.1214/18-AOS1777.
[26] G. Shen, Y. Jiao, Y. Lin, J. L. Horowitz, and J. Huang, “Nonparametric estimation of non-crossing quantile regression process with deep ReQU neural networks,” Journal of Machine Learning Research, vol. 25, no. 88, pp. 1–75, 2024.
[27] R. Koenker, “quantreg: Quantile Regression,” R package version 6.1, 2025,
