Transformations for Parametric Modeling of Censored Data in Survival Analysis

摘要

Some specific transformation functions are introduced in this paper to transform probability distributions commonly used for parametric modeling in survival analysis in order to make the analysis more flexible and prevent possible overestimation in longer survival times. The developed distributions were used to model the survival functions for two different data sets that contain information about patients suffering from breast cancer and colorectal carcinoma. The results are compared with the standard approach using Akaike's information criterion.