EUVIMEDEUVIMED entdecken
Crossref

Novel heteroscedastic robust ridge M-estimators for linear regression model

Hina Naz, Ismail Shah, Danish Wasim, Sajid Ali

Hacettepe Journal of Mathematics and Statistics · 02.02.2026 · (Advanced Online Publication)

DOI: 10.15672/hujms.1734907 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Hacettepe Journal of Mathematics and Statistics, online erschienen am 02.02.2026. Verfasst von Hina Naz und 3 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 1 Literaturangaben. Gefördert von: NA.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Penalized regression techniques are widely used when the number of predictors is large and they are highly inter-correlated. Within penalized linear regression techniques, ridge regression is primarily used to obtain estimators that minimize the mean squared error while fitting the regression model. This research work addresses the issue of ridge estimation when multicollinearity, heteroscedasticity, and outliers exist within the predictors. For example, in a dataset that predicts livestock production based on factors such as animal population and feed quality, an outlier, such as an unusually high number of livestock in one region due to a large farm, could distort the regression results. To this end, we introduced a novel heteroscedastic ridge M-estimator approach in conjunction with the scaling factor method. This new approach improves the performance of ridge estimators compared to traditional techniques. To demonstrate its effectiveness, the proposed methodology is compared with several widely used and latest ridge estimators. Extensive simulations demonstrate the remarkable performance of the proposed approach, particularly in the presence of severe multicollinearity, heteroscedasticity, and outliers. Furthermore, real world applications validate the potential of this methodology as a valuable tool for data analysis in scenarios characterized by collinear predictors and heteroscedastic errors.

Förderung

  • NA

Zitieren

APA

Naz, H., Shah, I., Wasim, D., & Ali, S. (2026). Novel heteroscedastic robust ridge M-estimators for linear regression model. Hacettepe Journal of Mathematics and Statistics. https://doi.org/10.15672/hujms.1734907

Vancouver

Naz H, Shah I, Wasim D, Ali S. Novel heteroscedastic robust ridge M-estimators for linear regression model. Hacettepe Journal of Mathematics and Statistics. 2026. doi:10.15672/hujms.1734907

BibTeX
@article{Naz2026,
  author = {Naz, Hina and Shah, Ismail and Wasim, Danish and Ali, Sajid},
  title = {Novel heteroscedastic robust ridge M-estimators for linear regression model},
  journal = {Hacettepe Journal of Mathematics and Statistics},
  year = {2026},
  number = {Advanced Online Publication},
  doi = {10.15672/hujms.1734907}
}
Kostenlos starten →