Some inferential techniques in survival analysis

Abstract

Survival analysis plays an important role in Statistics, as sometimes we have to deal with incomplete observations that are referred to as censored observations. This study focuses on dealing with randomly censored data as it frequently arise in medical research, epidemiological studies and various other fields. Nonparametric tests have the advantage of making minimal assumptions about the distributional forms of the population to be examined. Therefore, some nonparametric tests are developed for testing the hypothesis associated with location/scale parameters, when the data is randomly censored. These tests are applicable to compare the survival conditions in two or more underlying populations having incomplete observations (or censored observations). Critical values of these tests are obtained and the statistical powers of the proposed tests are also studied using Monte Carlo simulation study. Efficiency comparison is carried out with existing tests. Some real-life examples are illustrated to demonstrate the practical application of proposed tests. newline