cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar
mikes1098
Level I

Calculate Exact Confidence Interval Numbers

Sorry for the basic questions -- trying to figure out JMP.  I have scoured the manuals but can't seem to find the answer.

I have the fit below -- with the 95% individual confidence intervals turned on.  I am trying to calculate the exact numbers, i.e. @ Y=.60 what is the lower line X value.... it is close to X=.005.   I need to be able to calculate it for any Y value -- similar to how the inverse predictor works.  Please let me know if you need any more info...

Thanks in advance

3251_ScreenShot2886.jpg

1 ACCEPTED SOLUTION

Accepted Solutions
mpb
mpb
Level VII

Re: Calculate Exact Confidence Interval Numbers

1. Instead of Fit Y by X use Fit Model. In the Fit Model dialog you can specify the Log transform for "y axis" and "x axis" using the "Transform" red triangle pull down.

2. In the output window use the red triangle pull down next to "Response Log(y axis)" and choose Save Columns >  "Indiv Confidence Limit Formula". This places formulas for the upper and lower 95% CL for Individuals in the table.

3. In JMP using the menus choose Graph > Profiler.

4. In the Profiler select the lower CL formula for "Y, Prediction Formula" and click OK.

5. In the resulting window click the red triangle pull down next to "Prediction Profiler" and choose "Desirability Functions"

6. Alt click the Desirability Function and change it to "Match Target" and then set the middle value to be your target (0.6 in your example

7. Click the red triangle pull down next to "Prediction Profiler" and choose Maximize Desirability

In your example you should be able to see the result (0.0048582) on the horizontal axis label and the achieved target (0.6) on the vertical axis.

View solution in original post

2 REPLIES 2
mpb
mpb
Level VII

Re: Calculate Exact Confidence Interval Numbers

1. Instead of Fit Y by X use Fit Model. In the Fit Model dialog you can specify the Log transform for "y axis" and "x axis" using the "Transform" red triangle pull down.

2. In the output window use the red triangle pull down next to "Response Log(y axis)" and choose Save Columns >  "Indiv Confidence Limit Formula". This places formulas for the upper and lower 95% CL for Individuals in the table.

3. In JMP using the menus choose Graph > Profiler.

4. In the Profiler select the lower CL formula for "Y, Prediction Formula" and click OK.

5. In the resulting window click the red triangle pull down next to "Prediction Profiler" and choose "Desirability Functions"

6. Alt click the Desirability Function and change it to "Match Target" and then set the middle value to be your target (0.6 in your example

7. Click the red triangle pull down next to "Prediction Profiler" and choose Maximize Desirability

In your example you should be able to see the result (0.0048582) on the horizontal axis label and the achieved target (0.6) on the vertical axis.

mikes1098
Level I

Re: Calculate Exact Confidence Interval Numbers

Simply amazing... this not only answered my question, but gave me a bunch of little tips on navigating JMP.  I feel I am severely under utilizing this program.