×

A simulation study of ridge regression estimators with autocorrelated errors. (English) Zbl 0695.62176

Summary: We run a large number of simulations to study the effects of collinearity and autocorrelated disturbances in the performance of several ridge regression estimators. The results suggest that with a fair amount of multicollinearity and of autocorrelation the ridge regression estimators which take the autocorrelation into account can perform better than the other methods. Also if the error term is only moderately autocorrelated, then the performance of the ridge regression estimators built upon ignoring the autocorrelation can outperform the other estimators.

MSC:

62J07 Ridge regression; shrinkage estimators (Lasso)
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Gosling B.J., Communications in Statistics 14 pp 577– (1985) · Zbl 0576.62075 · doi:10.1080/03610918508812459
[2] DOI: 10.2307/2285435 · Zbl 0321.62071 · doi:10.2307/2285435
[3] DOI: 10.2307/1267351 · Zbl 0202.17205 · doi:10.2307/1267351
[4] DOI: 10.1080/03610927508827232 · Zbl 0296.62062 · doi:10.1080/03610927508827232
[5] DOI: 10.1080/03610927608827353 · Zbl 0336.62056 · doi:10.1080/03610927608827353
[6] DOI: 10.1080/03610928208828335 · Zbl 0506.62054 · doi:10.1080/03610928208828335
[7] Prais, S.J and Winsten, C.B. 1954. ”Trend Estimators and Serial Correlation”. Chicago Cowless Commision Discussion Paper N{\(\deg\)}383
[8] DOI: 10.1016/0304-4076(84)90045-9 · Zbl 0559.62054 · doi:10.1016/0304-4076(84)90045-9
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.