Estimation and diagnostic for single-index partially functional linear regression model with $ p $-order autoregressive skew-normal errors

Lijie Zhou; Liucang Wu; Bin Yang; Lijie Zhou; Liucang Wu; Bin Yang

doi:10.3934/math.2025321

AIMS Mathematics

2025, Volume 10, Issue 3: 7022-7066. doi: 10.3934/math.2025321

Previous Article Next Article

Research article Special Issues

Estimation and diagnostic for single-index partially functional linear regression model with $ p $-order autoregressive skew-normal errors

1.
Faculty of Science, Kunming University of Science and Technology, Kunming 650500, China
2.
City College, Kunming University of Science and Technology, Kunming 650500, China

Received: 08 December 2024 Revised: 10 March 2025 Accepted: 17 March 2025 Published: 26 March 2025
MSC : 62J05, 62M10

This paper introduced a novel single-index partially functional linear regression model with $ p $-order autoregressive skew-normal errors, addressing the dual challenges of autocorrelation and skewness in high-dimensional functional data. We proposed an innovative EM-CALS algorithm, which synergizes the expectation-maximization (EM) framework with conditional adaptive least squares to estimate constrained parameters and functional components efficiently. The algorithm uniquely addressed the unit-norm constraint of single-index vectors through reparameterization, thereby overcoming the limitations of conventional EM in nonlinear optimization. Comprehensive simulations demonstrated EM-CALS's superiority over the two-step iterative least squares (TSILS) algorithm, achieving reductions of 1.19% in root mean square error (RMSE), 2.45% in mean square error (MSE), 38.00% in root average squared errors ($ \text{RASE}_{1} $), and 51.35% in $ \text{RASE}_{2} $, respectively, demonstrating a clear advantage in enhancing prediction accuracy. Additionally, we conducted residual analysis based on conditional quantiles, considering the skew-normal distribution and autocorrelation of the residuals, and performed local influence analysis using the Q-function in the EM algorithm. The efficiency of the EM-CALS algorithm was validated through simulation studies. Finally, the methodology was applied to an empirical dataset from photovoltaic power forecasting, underscoring its practical applicability. This work bridged significant gaps in functional data analysis by simultaneously addressing dimension reduction, temporal dependence, and distributional asymmetry, which are crucial challenges in modern energy analytics and biomedical studies.
- single-index partially functional linear regression,
- B-spline,
- EM-CALS algorithm,
- autoregressive errors,
- skew-normal distribution,
- adaptive least squares,
- local inﬂuence
Citation: Lijie Zhou, Liucang Wu, Bin Yang. Estimation and diagnostic for single-index partially functional linear regression model with $ p $-order autoregressive skew-normal errors[J]. AIMS Mathematics, 2025, 10(3): 7022-7066. doi: 10.3934/math.2025321

Related Papers:

Abstract

This paper introduced a novel single-index partially functional linear regression model with $ p $-order autoregressive skew-normal errors, addressing the dual challenges of autocorrelation and skewness in high-dimensional functional data. We proposed an innovative EM-CALS algorithm, which synergizes the expectation-maximization (EM) framework with conditional adaptive least squares to estimate constrained parameters and functional components efficiently. The algorithm uniquely addressed the unit-norm constraint of single-index vectors through reparameterization, thereby overcoming the limitations of conventional EM in nonlinear optimization. Comprehensive simulations demonstrated EM-CALS's superiority over the two-step iterative least squares (TSILS) algorithm, achieving reductions of 1.19% in root mean square error (RMSE), 2.45% in mean square error (MSE), 38.00% in root average squared errors ($ \text{RASE}_{1} $), and 51.35% in $ \text{RASE}_{2} $, respectively, demonstrating a clear advantage in enhancing prediction accuracy. Additionally, we conducted residual analysis based on conditional quantiles, considering the skew-normal distribution and autocorrelation of the residuals, and performed local influence analysis using the Q-function in the EM algorithm. The efficiency of the EM-CALS algorithm was validated through simulation studies. Finally, the methodology was applied to an empirical dataset from photovoltaic power forecasting, underscoring its practical applicability. This work bridged significant gaps in functional data analysis by simultaneously addressing dimension reduction, temporal dependence, and distributional asymmetry, which are crucial challenges in modern energy analytics and biomedical studies.

References

[1]	H. Cardot, F. Ferraty, P. Sarda, Spline estimators for the functional linear model, Stat. Sin., 13 (2003), 571–591. Available from: http://www.jstor.org/stable/24307112.
[2]	D. Şentürk, H. G. Müller, Functional varying coefficient models for longitudinal data, J. Am. Stat. Assoc., 105 (2010), 1256–1264. https://doi.org/10.1198/jasa.2010.tm09228 doi: 10.1198/jasa.2010.tm09228
[3]	G. Aneiros-Pérez, P. Vieu, Semi-functional partial linear regression, Stat. Probab. Lett., 76 (2006), 1102–1110. https://doi.org/10.1016/j.spl.2005.12.007 doi: 10.1016/j.spl.2005.12.007
[4]	J. Zhou, M. Chen, Spline estimators for semi-functional linear model, Stat. Probab. Lett., 82 (2012), 505–513. https://doi.org/10.1016/j.spl.2011.11.027 doi: 10.1016/j.spl.2011.11.027
[5]	W. Härdle, P. Hall, H. Ichimura, Optimal smoothing in single-index Models, J. Comput. Graph. Stat., 21 (1993), 157–178. https://doi.org/10.1214/aos/1176349020 doi: 10.1214/aos/1176349020
[6]	Q. M. Zou, Z. Y. Zhu, M-estimators for single-index model using B-spline, Metrika, 77 (2014), 225–246. https://doi.org/10.1007/s00184-013-0434-z doi: 10.1007/s00184-013-0434-z
[7]	P. Yu, J. Du, Z. Zhang, Single-index partially functional linear regression model, Stat. Pap., 61 (2020), 1107–1123. https://doi.org/10.1007/s00362-018-0980-6 doi: 10.1007/s00362-018-0980-6
[8]	S. Dabo-Niang, S. Guillas, Functional semiparametric partially linear model with autoregressive errors, J. Multivar. Anal., 101 (2010), 307–315. https://doi.org/10.1016/j.jmva.2008.06.008 doi: 10.1016/j.jmva.2008.06.008
[9]	B. Yang, M. Chen, T. Su, J. Zhou, Robust estimation for semi-functional linear model with autoregressive errors, Mathematics, 11 (2023), 277. https://doi.org/10.3390/math11020277 doi: 10.3390/math11020277
[10]	P. Xiao, G. Wang, Partial functional linear regression with autoregressive errors, Commun. Stat. Theor.-M., 51 (2020), 4515–4536. https://doi.org/10.1080/03610926.2020.1818097 doi: 10.1080/03610926.2020.1818097
[11]	A. Azzalini, A class of distributions which includes the normal ones, Scand. J. Stat., 12 (1985), 171–178. Available from: http://www.jstor.org/stable/4615982.
[12]	J. Xiao, P. Yu, Z. Zhang, Weighted composite asymmetric Huber estimation for partial functional linear models, AIMS Math., 7 (2022), 7657–7684. https://doi.org/10.3934/math.2022430 doi: 10.3934/math.2022430
[13]	C. da Silva Ferreira, G. A. Paula, G. C. Lana, Estimation and diagnostic for partially linear models with first-order autoregressive skew-normal errors, Comput. Stat., 37 (2022), 445–468. https://doi.org/10.1007/s00180-021-01130-2 doi: 10.1007/s00180-021-01130-2
[14]	Y. Liu, J. Lu, G. A. Paula, S. Liu, Bayesian diagnostics in a partially linear model with first-order autoregressive skew-normal errors, Comput. Stat., 46 (2024), 1–31. https://doi.org/10.1007/s00180-024-01504-2 doi: 10.1007/s00180-024-01504-2
[15]	C. da Silva Ferreira, M. H. Montoril, G. A. Paula, Partially linear models with p-order autoregressive skew-normal errors, Braz. J. Probab. Stat., 36 (2022), 792–806. https://doi.org/10.1214/22-bjps556 doi: 10.1214/22-bjps556
[16]	L. Z. Chen, G. C. Wang, B. X. Zhang, Functional partial linear regression with autoregressive errors, Commun. Math. Stat., 2025,225–246. https://doi.org/10.1007/s40304-024-00409-2 doi: 10.1007/s40304-024-00409-2
[17]	M. Wang, M. L. Shu, J. J. Zhou, S. X. Wu, M. Chen, Least square estimation for multiple functional linear model with autoregressive errors, Acta Math. Appl. Sin. Engl. Ser., 41 (2025), 84–98. https://doi.org/10.1007/s10255-024-1143-2 doi: 10.1007/s10255-024-1143-2
[18]	R. D. Cook, Assessment of local influence, J. R. Stat. Soc. Ser. B, 48 (1986), 133–155. https://doi.org/10.1111/j.2517-6161.1986.tb01398.x doi: 10.1111/j.2517-6161.1986.tb01398.x
[19]	Q. M. Zou, Z. Y. Zhu, J. L. Wang, Local influence analysis for penalized Gaussian likelihood estimation in partially linear single-index models, Ann. Inst. Stat. Math., 61 (2009), 905–918. https://doi.org/10.1007/s10463-007-0158-9 doi: 10.1007/s10463-007-0158-9
[20]	S. K. Sahu, D. K. Dey, M. D. Branco, A new class of multivariate skew distributions with applications to bayesian regression models, Can. J. Stat., 31 (2003), 129–150. https://doi.org/10.2307/3316064 doi: 10.2307/3316064
[21]	C. de Boor, A practical guide to splines, New York: Springer, 2001. https://doi.org/10.2307/2006241
[22]	J. Huang, J. L. Horowitz, F. Wei, Variable selection in nonparametric additive models, Ann. Stat., 38 (2010), 2282–2313. https://doi.org/10.1214/09-aos781 doi: 10.1214/09-aos781
[23]	A. P. Dempster, N. M. Laird, D. B. Rubin, Maximum likelihood from incomplete data via the EM algorithm, J. Roy. Stat. Soc. B, 39 (1977), 1–38. https://doi.org/10.1111/j.2517-6161.1977.tb01600.x doi: 10.1111/j.2517-6161.1977.tb01600.x
[24]	N. L. Johnson, A. W. Kemp, S. Kotz, Univariate discrete distributions, New York: John Wiley, 2005. https://doi.org/10.1002/0471715816
[25]	K. Takai, Constrained EM algorithm with projection method, Comput. Stat., 27 (2012), 701–714. https://doi.org/10.1007/s00180-011-0285-x doi: 10.1007/s00180-011-0285-x
[26]	Y. Yu, D. Ruppert, Root-n consistency of penalized spline estimator for partially linear single-index models under general Euclidean space, Stat. Sin., 14 (2004), 449–456. Available from: https://api.semanticscholar.org/CorpusID: 14712886.
[27]	B. Yang, M. Chen, J. J. Zhou, Forecasting the monthly retail sales of electricity based on the semi-functional linear model with autoregressive errors, AIMS Math., 10 (2025), 1602–1627. https://doi.org/10.3934/math.2025074 doi: 10.3934/math.2025074
[28]	Q. Y. Peng, J. J. Zhou, N. S. Tang, Varying coefficient partially functional linear regression models, Stat. Pap., 57 (2015), 827–841. https://doi.org/10.1007/s00362-015-0681-3 doi: 10.1007/s00362-015-0681-3
[29]	H. Lin, J. Shi, T. Tong, R. Zhang, Partially linear single-index model in the presence of measurement error, J. Syst. Sci. Complex., 35 (2022), 2361–2380. https://doi.org/10.1007/s11424-022-1112-x doi: 10.1007/s11424-022-1112-x
[30]	P. Yu, Z. Zhang, J. Du, A test of linearity in partial functional linear regression, Metrika, 79 (2016), 953–969. https://doi.org/10.1007/s00184-016-0584-x doi: 10.1007/s00184-016-0584-x
[31]	P. K. Dunn, G. K. Smyth, Randomized quantile residuals, J. Comput. Graph. Stat., 5 (1996), 236–244. https://doi.org/10.2307/1390802 doi: 10.2307/1390802
[32]	G. Wang, Y. Su, L. Shu, One-day-ahead daily power forecasting of photovoltaic systems based on partial functional linear regression models, Renew. Energy, 96 (2016), 469–478. https://doi.org/10.1016/j.renene.2016.04.089 doi: 10.1016/j.renene.2016.04.089
[33]	H. T. Zhu, S. Lee, Local influence for incomplete data models, J. R. Stat. Soc. Ser. B, 63 (2001), 111–126. https://doi.org/10.1111/1467-9868.00279 doi: 10.1111/1467-9868.00279
[34]	Y. Qiu, Z. Zhou, Macao meteorological and geophysical Bureau, In: 2018-2019 PV Generation of a Rooftop Plant in the University of Macau and Weather Report, IEEE Dataport, 2020. https://dx.doi.org/10.21227/1khg-8t55

Reader Comments

Your name:*

Email:*
© 2025 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)