LOGISTIC RIDGE REGRESSION FOR CLINICAL DATA ANALYSIS (A CASE STUDY)

This paper focuses on regression with binomial response data. In these cases logit regression is the most used model. An example is a retrospective biomedical problem, where multicollinearity occurs, thus the variances of the estimated parameters are large. In this paper we propose to apply the ridge method to the maximum likelihood estimation of the logit model parameters. The efficiency of the proposed technique was investigated using a biomedical data set. A random sampling technique was used to study the effect of sample size on the ML and the logistic ML estimation.