Total stability of outcome weighted learning

Yuejia Sun; Dao-Hong Xiang

doi:10.3934/mfc.2024025

Article Contents

2025, Volume 8, Issue 5: 734-755. Doi: 10.3934/mfc.2024025 open access

This issue Previous Article Neutrosophic fuzzy $ I $-convergent Fibonacci difference sequence spaces Next Article Dissipative control for singular Takagi-Sugeno fuzzy systems with random actuator failures

Total stability of outcome weighted learning

Yuejia Sun^1, and
Dao-Hong Xiang^1, ,

1.
School of Mathematical Sciences, Zhejiang Normal University, Jinhua, Zhejiang 321004, China

^*Corresponding author: Dao-Hong Xiang
^*Corresponding author: Dao-Hong Xiang

Received: February 2024

Revised: April 2024

Early access: June 2024

Published: October 2025

D. H. Xiang is supported by Zhejiang Provincial Natural Science Foundation of China (Grant No.LY23A010009) and the National Natural Science Foundation of China (Grant No.U20A2068).

Abstract / Introduction Full Text(HTML) Related Papers Cited by

Abstract

The goal of personalized medicine is to identify the optimal treatment rule (ITR) for each patient to maximize each patient's expected clinical outcome by taking account into the patient's heterogeneity. Outcome weighted learning (OWL) is one of the most popular algorithms to estimate the optimal ITR. In this paper, we mainly studied the stability of the OWL algorithm, which investigated the influence of simultaneous small perturbations of the probability measure, regularization parameter, and the kernel. Traditional statistical robustness only considered the impact of slight changes in the probability measure. The results of this study were more generalized.

Keywords:

Mathematics Subject Classification: Primary: 62G35, 68Q32.

Citation:

FullText(HTML)

Related Papers

Cited by

References

[1]	N. Aronszajn, Theory of reproducing kernels, Transactions of the American Mathematical Society, 68 (1950), 337-404. doi: 10.2307/1990404.
[2]	A. Berlinet and C. Thomas-Agnan, Reproducing Kernel Hilbert Spaces in Probability and Statistics, With a preface by Persi Diaconis Kluwer Academic Publishers, Boston, MA, 2004. doi: 10.1007/978-1-4419-9096-9.
[3]	O. Bousquet and A. Elisseeff, Stability and generalization, Journal of Machine Learning Research, 2 (2002), 499-526.
[4]	O. Bousquet, Y. Klochkov and N. Zhivotovskiy, Sharper bounds for uniformly stable algorithms, Proceedings of Machine Learning Research, 125 (2020), 1-17.
[5]	A. Christmann and A. Van Messem, Bouligand derivatives and robustness of support vector machines for regression, Journal of Machine Learning Research, 9 (2008), 915-936.
[6]	A. Christmann, A. Van Messem and I. Steinwart, On consistency and robustness properties of support vector machines for heavy-tailed distributions, Statistics and its Interface, 2 (2009), 311-327. doi: 10.4310/SII.2009.v2.n3.a5.
[7]	A. Christmann and I. Steinwart, On robustness properties of convex risk minimization methods for pattern recognition, Journal of Machine Learning Research, 5 (2003/04), 1007-1034.
[8]	A. Christmann and I. Steinwart, Consistency and robustness of kernel based regression in convex risk minimization, Bernoulli, 13 (2007), 799-819. doi: 10.3150/07-BEJ5102.
[9]	A. Christmann, D. H. Xiang and D. X. Zhou, Total stability of kernel methods, Neurocomputing, 289 (2018), 101-118.
[10]	A. Christmann and D.-X. Zhou, On the robustness of regularized pairwise learning methods based on kernels, Journal of Complexity, 37 (2016), 1-33. doi: 10.1016/j.jco.2016.07.001.
[11]	F. Cucker and S. Smale, On the mathematical foundations of learning, Bulletin of the American Mathematical Society, 39 (2002), 1-49. doi: 10.1090/S0273-0979-01-00923-5.
[12]	F. Cucker and D. X. Zhou, Learning Theory: An Approximation Theory Viewpoint, Cambridge University Press, Cambridge, 2007. doi: 10.1017/CBO9780511618796.
[13]	V. Feldman, J. Vondrak and K. Sridharan, Generalization bounds foruniformly stable algorithms, In Advances in Neural Information Processing Systems, (2018), 9747-9757.
[14]	V. Feldman, J. Vondrak and K. Sridharan, High probability generalization bounds for uniformly stable algorithms with nearly optimal rate, In Conference on Learning Theory, (2019), 1270-1279.
[15]	D. J. Foster, S. Greenberg, S. Kale, H. Luo, M. Mohri and K. Sridharan, Hypothesis set stability and generalization, In Advances in Neural Information Processing Systems, (2019), 6726-6736.
[16]	W.-C. Guo, X.-H. Zhou and S.-J. Ma, Estimation of optimal using a covariate-specific treatment effect curve with high-dimensional covariates, Journal of the American Statistical Association, 116 (2021), 309-321. doi: 10.1080/01621459.2020.1865167.
[17]	R. Hable, Asymptotic normality of support vector machine variants and other regularized kernel methods, Journal of Multivariate Analysis, 106 (2012), 92-117. doi: 10.1016/j.jmva.2011.11.004.
[18]	R. Hable and A. Christmann, On qualitative robustness of support vector machines, Journal of Multivariate Analysis, 102 (2011), 993-1007. doi: 10.1016/j.jmva.2011.01.009.
[19]	H. Kohler and A. Christmann, Total stability of SVMs and localized SVMs, Journal of Machine Learning, 23 (2022), 1-41.
[20]	Y. Lei and Y. Ying, Fine-Grained Analysis of Stability and Generalization for Stochastic Gradient Descent, International Conference on Machine Learning, 2020.
[21]	T. Liu, G. Lugosi, G. Neu and D. Tao, Algorithmic stability and hypothesis complexity, In International Conference on Machine Learning, (2017), 2159-2167.
[22]	Y. Liu, Y.-J. Wang and M. R. Kosorok, Robust hybrid learning for estimating personalized dynamic treatment regimens, preprint, 2016, arXiv: 1611.02314.
[23]	Y. Liu, H. Zhang and Y. Wu, Hard or soft classification? Large-margin unified machines, Journal of the American Statistical Association, 106 (2011), 166-177. doi: 10.1198/jasa.2011.tm10319.
[24]	J. S. Marron, M. J. Todd and J. Ahn, Distance-weighted discrimination, Journal of the American Statistical Association, 102 (2007), 1267-1271. doi: 10.1198/016214507000001120.
[25]	S. Mukherjee, P. Niyogi, T. Poggio and R. Rifkin, Learning theory: Stability is sufficient for generalization and necessary and sufficient for consistency of empirical risk minimization, Advances in Computational Mathematics, 25 (2006), 161-193. doi: 10.1007/s10444-004-7634-z.
[26]	S. A. Murphy, Optimal dynamic treatment regimes, Journal of the Royal Statistical Society: Series B, 65 (2003), 331-355.
[27]	S. A. Murphy, A generalization error for Q-learning, Journal of Machine Learning Research, 6 (2005), 1073-1097.
[28]	Z.-L. Qi and Y.-F. Liu, D-learning to estimate optimal individual treatment rules, Electronic Journal of Statistics, 12 (2018), 3601-3638. doi: 10.1214/18-ejs1480.
[29]	M. Qian and S. A. Murphy, Performance guarantees for individualized treatment rules, Annals of Statistic, 39 (2011), 1180-1210. doi: 10.1214/10-AOS864.
[30]	K. S. Shah, H. D. Fu and M. R. Kosorok, Stabilized direct learning for efficient estimation of individualized treatment rules, Btometric Methodology, 79 (2023), 2843-2856. doi: 10.1111/biom.13818.
[31]	I. Steinwart and A. Christmann, Support Vector Machines, Springer, New York, 2008.
[32]	V. N. Vapnik, The Nature of Statistical Learning Theory, Springer, New York, 1995. doi: 10.1007/978-1-4757-2440-0.
[33]	V. N. Vapnik, Statistical Learning Theory, John Wiley and Sons, New York, 1998.
[34]	C. J. Watkins and P. Dayan, Q-learning, Machine Learning, 8 (1992), 279-292.
[35]	D. Werner, Funktionalanalysis, 4$^{th}$ edition, Springer, Berlin, 2002.
[36]	Y.-Q. Zhao and D.-L. Zeng, Estimating individualized treatment rules using outcome weighted learning, Journal of the American Statistical Association, 107 (2012), 1106-1118. doi: 10.1080/01621459.2012.695674.
[37]	X. Zhou, M. H. Nicole, K. Umer and M. R. Kosorok, Residual weighted learning for estimating individuaized treatment rules, Journal of the American Statistical Association, 112 (2017), 169-187. doi: 10.1080/01621459.2015.1093947.
[38]	X.-X. Zhu, Error Analysis of OWL Algorithms with Varying Gaussians and Convex Loss, Master thesis, Zhejiang Normal University in Jin Hua, 2022.