Non-parametric Survival Techniques with Improved Weight

Abstract

Generally, in most of the studies all the subjects could not attain the event of interest within the time period, which results in censoring, also sometimes leads to the situation of heavy censoring. In the case of heavy censoring, the conventional Kaplan-Meier estimator overestimates the survival function. To overcome this, a new weight is designed by accounting the length of the survival time and censoring rate for each time interval and is evaluated from the existing dataset for various amount of censoring percentage. Further, to validate the new estimates of survival function is better, some measures like standard error and confidence intervals are obtained empirically. Finally, to make the computation easier, coding has been written using R-Software.

Keywords: Censoring, survival function, Kaplan-Meier estimate, median survival time, standard error and confidence-interval

Cite this Article

M. Ramadurai, M.A. Ghouse Basha. Non-parametric Survival Techniques with Improved Weight. Research & Reviews: A Journal of Bioinformatics. 2018; 5(3):
19–29p.