Non-Parametric Methods of Survival Analysis

Techniques of survival analysis play increasing important roles in Biostatistics, modern medical research, engineering, marketing and demography among others. Researchers carry out survival analysis, in mentioned areas, on data such as relapse of a disease, duration of the failure of a machine, survival of a new business, duration of the marriage, time to death, the length of stay in jail/hospital, the occurrence of a strike or of a risk insured against, the length of time people remain unemployed after a job loss or after graduating from training among others. Due to considerable methodological development concerning this statistical area, today, methods of survival analysis exist in three-fold: parametric, semi-parametric and non-parametric. However, the non-parametric method is the most used technique due to relaxing conditions about the assumptions of the distributions. Several non-parametric methods have been considered in this book; it is observed that the performance of Adjusted Product Limit Estimator is quite satisfactory.