CAGE User's Guide

Optimizing the Feature

After filling the feature, you can improve the fit of the strategy to the model by optimizing the feature. The optimization routine does the following:

• First CAGE optimizes the breakpoints for the normalizers. (See Optimizing Breakpoints.)
• Next CAGE optimizes the values of the tables. (See Optimizing Table Values.)

This section gives the procedure for optimizing a feature. For further details of how optimization works, see the references given above.

To optimize a feature:

1. Click . This opens the Feature Optimizing Options dialog box, as shown.

   

2. Enter the ranges of the normalizers for the normalizer optimization.
3. Enter the numbers of grid points for the normalizers.
4. Enter the values for the other variables.
5. Enter the ranges of the normalizers for the table optimization.
6. Enter the numbers of grid points for the normalizers.
7. Enter the values of the other variables.
8. Enter the table fill order.

1. In this example, enter [1 2 1]. This optimizes the normalizers and then the table values of Table_NL. Next CAGE optimizes the normalizers, and then the table values of table Fn_A. After which CAGE optimizes the normalizers and then the table values for Table_NL again.

9. Click OK.

Note    You can iterate this process using the table fill order edit box, as in the example. This further improves the fit of the strategy and the model.

When you have completed a calibration, you can export your feature. For information, see Exporting Calibrations.

Filling the Feature

Feature View