The Logistic Regression Model with a Modified Weight Function in Survival Analysis

Abstract

Most of the Ridge regression estimators can only achieve one property or the other, namely, variance reduction, bias reduction or reduced Mean Square Error. To achieve both variance and bias reduction in Logistic Ridge regression the Modified Logistic Ridge regression estimator is designed. The estimator is used to model the survival function of diabetic patients who are exposed to some specified medication. The model is formulated in such a way that the response probability is made to act as survival function. By some radical exponentiation of the weight function, the proposed estimator is found to have smaller bias than the Generalized Ordinary Logistic Ridge estimator.

Keywords: logistic ridge estimator, survival function, response probability, collinearity, means square error, bias