Journal of Statistical and Econometric Methods

A New Logistic Ridge Regression Estimator Using Exponentiated Response Function

• [ Download ]
• Times downloaded: 10889

• Abstract

  The logistic ridge regression estimator was designed to address the problem of variance inflation created by the existence of collinearity among the explanatory variables in logistic regression models. To reduce the bias introduced by the logistic ridge estimator, and at the same time achieve variance reduction, a modified generalized logistic ridge regression estimator is proposed in this paper. By exponentiating the response function, the weight matrix is enhanced, thereby reducing the variance. The modified estimator is jackknifed to achieve bias reduction.