Lognormal MLE Solution with Right Censored Data

From ReliaWiki
Jump to navigation Jump to search
Weibull Reference Examples Banner.png


New format available! This reference is now available in a new format that offers faster page load, improved display for calculations and images and more targeted search.

As of January 2024, this Reliawiki page will not continue to be updated. Please update all links and bookmarks to the latest references at Weibull examples and Weibull reference examples.




Lognormal MLE Solution with Right Censored Data

This example validates the calculations for the Lognormal MLE solution with Fisher matrix bound for right censored data in Weibull++ standard folios.


Reference Case

The data set on page 199 of the book Statistical Methods for Reliability Data by Dr. Meeker and Dr. Escobar, John Wiley & Sons, 1998 is used.


Data

Number in State State F or S Time to Failure
1 F 252
1 F 315
1 F 369
1 F 403
1 F 474
15 S 500


Result

  • The MLE solution is [math]\displaystyle{ \hat{\mu} = 6.56, \ \hat{\sigma} = 0.543\,\! }[/math].
  • The variance and covariance matrix is
[math]\displaystyle{ \sum =\begin{bmatrix} 0.0581 & 0.0374 \\ 0.0374 & 0.0405 \end{bmatrix}\,\! }[/math]


Results in Weibull++


  • The following picture shows the MLE solution and the variance/covariance matrix:
Lg right censored matrix.png


  • The Fisher matrix bound for parameters are:
For Ln-mu (using normal approximation of Eqn. 8.7 on page 187):
[math]\displaystyle{ [\hat{\mu}'_{L}, \hat{\mu}'_{U}] = \hat{\mu}' \pm z_{(1-\alpha / 2)}se_{\hat{\mu}'}\,\! }[/math]


For a confidence level of 0.95, it is:
[math]\displaystyle{ \begin{alignat}{2} [\hat{\mu}'_{L}, \hat{\mu}'_{U}] =& \hat{\mu}' \pm z_{(1-\alpha /2)}se_{\hat{\mu}'}\\ =& 6.564256 \pm 1.96 \times (0.0581)^{0.5}\\ =& [6.0918, 7.0366933]\\ \end{alignat}\,\! }[/math]


For Ln-Std (using log-normal approximation of Eqn. 8.8 on page 188):
[math]\displaystyle{ [\hat{\sigma}'_{L}, \hat{\sigma}'_{U}] = \hat{\sigma}'exp (\pm \frac{z_{(1-\alpha /2)}se_{\hat{\sigma}'}}{\hat{\sigma}'})\,\! }[/math]


For confidence level of 0.95, it is:
[math]\displaystyle{ \begin{alignat}{2} [\hat{\sigma}'_{L}, \hat{\sigma}'_{U}] =& \hat{\sigma}'exp (\pm \frac{z_{(1-\alpha /2)}se_{\hat{\sigma}'}}{\hat{\sigma}'})\\ =& 0.534049 \times exp(\pm \tfrac{1.96\times 0.040562^{0.5}}{0.534049})\\ =& [0.255, 1.118]\\ \end{alignat} \,\! }[/math]


The following picture shows the results in Weibull++:
Lg right censored bounds.png