The Performance of Quantile Regression and Linear Regression with Heteroskedasticity was Compared in a Simulated Study

Authors

Marwa R. Saad, Department of Applied Statistics and Econometrics, Faculty of Graduate studies for Statistical Research, Cairo University, Cairo, EgyptFollow
Ahmed H. Youssef, Department of Applied Statistics and Econometrics, Faculty of Graduate studies for Statistical Research, Cairo University, Cairo, EgyptFollow
Shereen H. Abdel Latif, Department of Applied Statistics and Econometrics, Faculty of Graduate studies for Statistical Research, Cairo University, Cairo, EgyptFollow

Author Country (or Countries)

Egypt

Abstract

The least-square estimator has several drawbacks when dealing with heteroscedasticity; this estimate will not be a Best Linear Unbiased Estimator (BLUE). Quantile Regression is a dependable option; however, it has some substantial computational problems. We compare five resampling approaches to estimate the standard error of the coefficients, in the situation of heterogeneity, for inference. According to simulation study, quantile regression beats linear regression and is also better when predicting errors in the presence of heterogeneity.

Digital Object Identifier (DOI)

http://dx.doi.org/10.18576/jsap/120226

Recommended Citation

R. Saad, Marwa; H. Youssef, Ahmed; and H. Abdel Latif, Shereen (2023) "The Performance of Quantile Regression and Linear Regression with Heteroskedasticity was Compared in a Simulated Study," Journal of Statistics Applications & Probability: Vol. 12: Iss. 2, Article 27.
DOI: http://dx.doi.org/10.18576/jsap/120226
Available at: https://digitalcommons.aaru.edu.jo/jsap/vol12/iss2/27