Performance of Ridge Estimators Based on Weighted Geometric Mean and Harmonic Mean

Autor:	Satish S. Bhat, R. Vidya
Rok vydání:	2020
Předmět:	Harmonic mean Estimator Weighted geometric mean Ridge (differential geometry) Geodesy Mathematics
Zdroj:	Journal of Scientific Research. 12:1-13
ISSN:	2070-0245 2070-0237
Popis:	Ordinary least squares estimator (OLS) becomes unstable if there is a linear dependence between any two predictors. When such situation arises ridge estimator will yield more stable estimates to the regression coefficients than OLS estimator. Here we suggest two modified ridge estimators based on weights, where weights being the first two largest eigen values. We compare their MSE with some of the existing ridge estimators which are defined in the literature. Performance of the suggested estimators is evaluated empirically for a wide range of degree of multicollinearity. Simulation study indicates that the performance of the suggested estimators is slightly better and more stable with respect to degree of multicollinearity, sample size, and error variance.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::68e757fe9c0c0f45032d7236a33019c6 https://doi.org/10.3329/jsr.v12i1.40525 Zobrazit plný text záznamu Plný text ve formátu PDF