The Alternate Cut-off Confusion Matrix JMP Add-In is designed to allow the user to specify a cut-off value or a range of values to generate a new confusion matrix for a binary response variable.
Upon launching the Alternate Cut-off Confusion Matrix JMP Add-In, you will be prompted for the following options:
Target: This column must have a nominal modeling type and contain only two levels.
Predicted Value: This column must have a continuous modeling type and should be one of the predicted (or probability) columns saved from the Partition, Neural (JMP Pro only), or Fit Model report.
Validation: This column must be numeric and should only be used if a validation column was used in the analysis. If a validation column was used in the analysis, the same validation column should also be specified here. Validation is only available in JMP Pro.
Alternate Cut-off Value: You can specify a cut-off value by typing the value in the dialog.
IMPORTANT NOTE: You must click the buttons to cast columns into the Target, Predicted Value,
and Validation roles. Dragging and dropping may result in an error message when you click OK.
Upon clicking OK, a new window will appear containing the desired confusion matrix and confusion rates. The launch dialog will remain open so that you can enter additional cut-off values.