Share your ideas for the JMP Scripting Unsession at Discovery Summit by September 17th. We hope to see you there!
Choose Language Hide Translation Bar
Highlighted
AlexVishay
Level I

Reliability distribution upper and lower parameter estimate

surv.jpg

  • I measure 6 events of failure and calculate, with JMP reliability function, the Kaplan-Meier survival function and Weibull fit (red curves).
  • I plot also lower 95% and upper 95% function (pink curves).
  • I repeat the analysis adding 6 measurements (total of 12) and I get the blue curves and upper, lower 95% (light blue).
  • Questions:

1.How lower 95% and upper 95% are calculated? Maximum likelihood estimator ratio???

2.I have only plotted the lower 95% alpha lower 95% beta and upper 95% alpha upper 95% beta curves, but I can also plot the lower 95% alpha upper 95% beta and upper 95% alpha lower 95% beta curves, do these combinations of parameters also have a meaning in defining the uncertainty interval of my distribution?

 

Thanks in advance for answers

5 REPLIES 5
Highlighted

Re: Reliability distribution upper and lower parameter estimate

1. JMP Help says:

 

"The Life Distribution platform offers two methods for calculating confidence intervals for the distribution parameters. These methods are labeled as Wald or Likelihood and can be selected in the launch window for the Life Distribution platform. Wald confidence intervals are used as the default setting. The computations for the confidence intervals for the cumulative distribution function (cdf) start with Wald confidence intervals on the standardized variable. Next, the intervals are transformed to the cdf scale (Nelson 1982, pp. 332–333 and pp. 346-347). The confidence intervals given in the other graphs and profilers are transformed Wald intervals (Meeker and Escobar 1998, ch. 7). Joint confidence intervals for the parameters of a two-parameter distribution are shown in the log-likelihood contour plots. They are based on approximate likelihood ratios for the parameters (Meeker and Escobar 1998, ch. 8)."

 

See Help > Books > Reliability and Survival.

Learn it once, use it forever!
Highlighted

Re: Reliability distribution upper and lower parameter estimate

2. The parameters do not have curves. They are presented as point estimates and interval estimates. My first reply describes how the intervals are calculated.

 

See Meeker and Escobar book.

Learn it once, use it forever!
Highlighted
txnelson
Super User

Re: Reliability distribution upper and lower parameter estimate

You may be able to get your questions answered by selecting the "?" tool, and then clicking on the output you are questioning.  It will bring up help screen(s) that will describe the output.  Or, you may find in in the book on reliability that was downloaded when you installed JMP:

     Help==>Books==>Reliability and Survival Methods

Jim
Highlighted
AlexVishay
Level I

Re: Reliability distribution upper and lower parameter estimate

For the moment, I use version 5 of JMP, so I think that help function is not avaiable with the click.

I see help and manual but, at least in version 5 manuals, I cannot find reference to Wald or Likelihood estimate.

 

Thank for help.

Highlighted

Re: Reliability distribution upper and lower parameter estimate

I can't remember any particular details about JMP 5. This version does not include the newer reliability and survival platforms or the updates to the older platforms with likelihood profile intervals. You should have Analyze > Fit Model, then Personality > Parametric Survival.

Learn it once, use it forever!
Article Labels

    There are no labels assigned to this post.