A Comparison of Semi-Parametric and Nonparametric Methods for Estimating Mean Time to Event for Randomly Left Censored Data

Authors

Farzana Chowdhury, Department of Business Administration, Northern University BangladeshFollow
Jahida Gulshan, Institute of Statistical Research and Training, University of DhakaFollow
Syed Shahadat Hossain, Institute of Statistical Research and Training, University of DhakaFollow

Abstract

The aim of this study was to make a comparison among existing estimation methods (Kaplan-Meier, Nelson-Aalen and Regression on Ordered Statistics (ROS)) for randomly left censored time to event data under selected distributions and for different level of censoring and sample sizes in order to determine the strength of these methods based on simulated data. Comparisons among the methods are made on the basis of unbiasedness and Monte Carlo Standard Error of the summary statistics (mean time to event) obtained by those methods under different conditions.

DOI

10.22237/jmasm/1430453760