We will explain this using the JMP sample data "Ingots.jmp".
1. In Fit Model, run a nominal logistic fit with the following options and select ROC Curve from the report options:
Condition: Y
Heating time, soaking time: Model effect configuration
Frequency: Frequency
2. Look at the ROC table and confirm that the probability of the row with an asterisk (*) is 0.9640.
3. Next, select Save Probability Formula from the red triangle button ( ▼ ) in the upper left corner of the report.
4. Several columns will be added to the end of the data table. Look at the column for Probability [Ready] (probability of a positive level).
From these, find the row that matches the probability "0.9640" shown above (it is rounded, so look for 0.96399...), and the values of "Heating time" and "Soaking time" in that row will become the cutoff values. In this example, "Heating time = 27" and "Soaking time = 1.0".
FAQ #3703
