Difference between JMP and reliasoft for calculating a 2P weibull

Question

A JMP user has received results from other teams doing reliability with reliasoft. His wish is to validate that JMP could achieve the same results, or even better.

I am interested in your opinion on the difference observed between JMP and the competition soft on the 2-parameter Weibull plot. The values indicated by the competition soft are included in the 95% confidence interval given by JMP but he would like to try to understand the origin of the difference. I have not identified any difference on the public formulas published in the user guides, neither in the algorithm used.

Here the results with Reliasoft:

With JMP, I have the following results

Nota: no Censored data

Thanks a lot for your feddbacks

Florence

DonMcCormack · Answer

Estimating parameters for a 2P Weibull is pretty standard. If Reliasoft is using maximum likelihood, the values should match. Are they using MLE? Is it the point estimate that's getting reported? Are the LEG values modeled separately or together (as you might with Fit Life by X)?

peng_liu · Answer

Reliasoft results could be from one of the methods different from our likelihood based estimation method.