×

On difference-based variance estimation in nonparametric regression when the covariate is high dimensional. (English) Zbl 1060.62047

Summary: We consider the problem of estimating the noise variance in homoscedastic non-parametric regression models. For low dimensional covariates \(t\in\mathbb{R}^d\), \(d=1,2\), difference-based estimators have been investigated in a series of papers. For a given length of such an estimator, difference schemes which minimize the asymptotic mean-squared error can be computed for \(d=1\) and \(d=2\). However, from numerical studies it is known that for finite sample sizes the performance of these estimators may be deficient owing to a large finite sample bias.
We provide theoretical support for these findings. In particular, we show that with increasing dimension \(d\) this becomes more drastic. If \(d\geq 4\), these estimators even fail to be consistent. A different class of estimators is discussed which allow better control of the bias and remain consistent when \(d\geq 4\). These estimators are compared numerically with kernel type estimators (which are asymptotically efficient), and some guidance is given about when their use becomes necessary.

MSC:

62G08 Nonparametric regression and quantile regression
65C20 Probabilistic models, generic numerical methods in probability and statistics

Software:

spatial; KernSmooth; NAG
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Bissantz N., Mnthly Notes R. Astron. Soc. 336 pp 131– (2002)
[2] Carroll R., Transforming and Weighting in Regression (1988) · doi:10.1007/978-1-4899-2873-3
[3] DOI: 10.1111/1467-9868.00152 · Zbl 0944.62041 · doi:10.1111/1467-9868.00152
[4] Fan J., Biometrika 85 pp 645– (1998)
[5] Gasser T., Biometrika 73 pp 625– (1986)
[6] Hall P., J. R. Statist. Soc. 51 pp 3– (1989)
[7] Hall P., Biometrika 77 pp 521– (1990) · Zbl 1033.62031
[8] Hall P., Adv. Appl. Probab. 23 pp 476– (1991)
[9] Hall P., Biometrika 77 pp 415– (1990) · Zbl 1033.62031
[10] DOI: 10.1016/S0304-4076(97)00044-4 · Zbl 0904.62047 · doi:10.1016/S0304-4076(97)00044-4
[11] Herrmann E., J. R. Statist. Soc. 57 pp 171– (1995)
[12] Jong P., Probab. Theory Reltd Flds 75 pp 261– (1987)
[13] Kay J., Lect. Notes Comput. Sci. 301 pp 587– (1988) · doi:10.1007/3-540-19036-8_59
[14] Lee J. S., Comput. Graph. Image Process. 15 pp 380– (1981)
[15] Muller H., Ann. Statist. 15 pp 610– (1987)
[16] DOI: 10.1111/1467-9469.00303 · Zbl 1035.62063 · doi:10.1111/1467-9469.00303
[17] Neumann J., Ann. Math. Statist. 12 pp 367– (1941)
[18] Neumann M., Statistics 25 pp 189– (1994)
[19] Numerical Algorithms Group, NAG C Library, Mark 5 (1998)
[20] Rice J., Ann. Statist. 12 pp 1215– (1984)
[21] Ripley B., Spatial Statistics (1981) · Zbl 0583.62087 · doi:10.1002/0471725218
[22] Ruppert D., Technometrics 39 pp 262– (1997)
[23] Seifert B., Biometrika 80 pp 373– (1993)
[24] Spokoiny V., J. Multiv. Anal. 82 pp 111– (2002)
[25] Stoer J., Einfuhrung in die Numerische Mathematik (1979) · doi:10.1007/978-3-662-06863-2
[26] Thompson A., Biometrika 78 pp 475– (1991)
[27] T. Wagner (1999 ) Untersuchungen uber die Varianz in nichtparametrischer Regression.PhD Thesis. Fakultat fur Mathematik, Ruhr-Universtat Bochum, Bochum. · Zbl 0998.62508
[28] Wand M., Kernel Smoothing (1995) · doi:10.1007/978-1-4899-4493-1
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.