Comparative Studies of the Reliability Estimation for the Mixed Weibull Distribution

• 발행기관: 한국자료분석학회
• 저자명: Hee Jae Kim
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.11 No.1, 63~72쪽, 전체 10쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2009.02.28

영문 초록

The Weibull distribution has been recognized as a useful model for survival populations associated with reliability studies and life testing experiments. This paper deals with estimating reliability from a mixture of two Weibull distributions. First, the maximum likelihood estimation is used to estimate the reliability of the mixed model when data are grouped and Type-I censored. Second, the Bayesian estimation is used to estimate the reliability. Third, the plug-in estimation is used to estimate the reliability. Each estimation will be compared numerically in the terms of various mission times, parameter values in the sense of asymptotic relative efficiency through Monte Carlo simulation.

목차

1. Introduction

2. The Mixture Model of Two Weibull Distributions

3. The Method of Reliability Estimation

4. Numerical Examples and Conclusion

Reference

키워드

Mixed Weibull distribution Mixed proportion Maximum likelihood Bayesian estimation Plug-in estimation Asymptotic relative efficiency Monte Carlo simulation Nonlinear simultaneous equation