Related Experiment Videos
Inference for reliability and stress-strength for a scaled Burr type X distribution
1Department of Mathematics and Statistics, Texas Tech University, Lubbock, Texas 79409, USA. surles@math.ttu.edu
Lifetime Data Analysis
|July 19, 2001
Summary
This study develops new statistical methods for reliability analysis when data follows a Burr type X distribution. These techniques help estimate the probability of failure (R) in engineering applications.
Area of Science:
- Statistics
- Probability Theory
- Reliability Engineering
Background:
- Reliability analysis is crucial in engineering.
- The Burr type X distribution is used for modeling.
- Exact inference for reliability (R) is often challenging.
Purpose of the Study:
- To develop statistical inference procedures for R = P(Y < X) under the Burr type X model.
- To address the lack of exact inference methods for this probability.
- To provide practical tools for reliability estimation.
Main Methods:
- Derivation of the expected Fisher information matrix.
- Development of asymptotic inference procedures for R and general functions of parameters.
- Application of a bootstrap method for estimating variance of maximum likelihood estimators.
Main Results:
- Asymptotic inference procedures for R were successfully developed.
- The expected Fisher information matrix was derived.
- Bootstrap method proposed for variance estimation.
Conclusions:
- The developed asymptotic and bootstrap methods offer effective solutions for reliability inference.
- These techniques are applicable to real-world data, such as carbon fiber strength.
- Simulations confirm the practical utility of the proposed methods.