You can calculate
mining specific quality measures by using the
functions DM_getConfMatrix and DM_getRMSE.
The following
quality measures are available:
- The table function DM_getConfMatrix holds
the confusion
matrix and cost matrix information contained in a value of type DM_ClasTestResult or DM_ClasModel.
If the value is of type DM_ClasModel, the confusion
and cost
matrix are returned from the validation result.
- For values
of type DM_RegressionModel or DM_RegTestResult,
the function DM_getRMSE returns the root mean squared
error
between actual and predicted target field values on the test data.
It returns
a value in absolute units (comparable to the original range in the
target
field) to describe the expected deviation of a predicted value from
the actual
value.