cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
The Discovery Summit 2025 Call for Content is open! Submit an abstract today to present at our premier analytics conference.
Choose Language Hide Translation Bar
View Original Published Thread

Tolerance intervals method

SengYew
Level II

Dear sir/Madam,
As we know that is only have two available methods are Assume Normal Distribution and Nonparametric from tolerance intervals module, If I have 30 samples data that using Fitted Weibull/lognormal distribution and the P-value is >0.05. How do we using Tolerance interval for Weibull and lognormal distribution?

3 REPLIES 3
MRB3855
Super User


Re: Tolerance intervals method

Hi @SengYew : If your data are Lognormal, then the natural log of the data will be normally distributed.

1. If X is your data, calculate Y=ln(X): Y has a normal distribution.

2. Get the desired Normal Tolerance Interval for Y.

3. Back transform the interval ( exp[lower TI], exp[upper TI] ) to get the Tolerance interval back on the original X scale.

 

Weibull? Not so straightforward I'm afraid:  

See the following: Lawless, J. F. (1975). Construction of tolerance bounds for the extreme-value and the Weibull distribution. Technometrics, 17, 255—261.

scappuccio
Level II


Re: Tolerance intervals method

Hi @MRB3855,

 

Based on your above response I am assuming the answer is no - does JMP have integrated tools for construction of Weibull tolerance intervals per the paper you have referenced?

MRB3855
Super User


Re: Tolerance intervals method

Hi @scappuccio   You assume correctly; the answer is no. However, I don’t know if there are plans to add that functionality in the future.