Comparison of nonparametric estimators versus parametric for reliability function

One of the main objectives of the area of realiability is to estimate the function of reliability, where traditionally are used non parametric estimators, being more efficient in sample of big sizes. In this work nonparametric estimators are compared to the reliability function through the mean squa...

全面介绍

书目详细资料
主要作者: Ramírez-Montoya, Javier
格式: Online
语言:spa
出版: Universidad Pedagógica y Tecnológica de Colombia - UPTC 2015
主题:
在线阅读:https://revistas.uptc.edu.co/index.php/ingenieria_sogamoso/article/view/4246
实物特征
总结:One of the main objectives of the area of realiability is to estimate the function of reliability, where traditionally are used non parametric estimators, being more efficient in sample of big sizes. In this work nonparametric estimators are compared to the reliability function through the mean square error using nonparametric estimators of Kaplan & Meier (1958), Nelson estimator (1969) and Bootstrap applied to Kaplan & Meier and Nelson. The comparison is made considering the parametric estimates, through simulation with different scenarios, times of interest, sizes sample and percentages of censorship, showing that the Bootstrap resampling normal type does not present the best results with Kaplan & Meier. Using Nelson, the 18% was more efficient.