Principal simple linear regression

Authors : Heydar Ali Mardanifard

Pages : 524-536

Doi:10.15672/hujms.1303627

View : 192 | Download : 151

Publication Date : 2024-04-23

Article Type : Research Paper

Abstract :In this paper, we propose a new method called principal simple linear regression for predicting a continuous response variable, $Y$ using a single continuous predictor, $X$ by utilizing multiple regression lines. This method is based on the theory of principal points and offers several advantages over classical simple linear regression methods, such as the ability to predict central, dispersion, and distributional tendencies of $Y$ on $X$, and simultaneous estimation. We provide the main properties, inferences, and limiting behavior of the estimators. Additionally, we conduct a comprehensive simulation study to validate our theoretical results. The model is also applied to real datasets to demonstrate its effectiveness.
Keywords : Simple linear regression, principal points, quantile regression