The CLAS_COST_TABLE_NAME setting specifies the name of a cost matrix table to be used in building a Decision Tree model. A cost matrix biases a classification model to minimize costly misclassifications. The cost matrix table must have the columns shown in Table 5-13.
Table 5-13 Cost Matrix Table Required Columns
| Column Name | Data Type |
|---|---|
|
|
valid target data type |
|
|
valid target data type |
|
|
|
Note: Valid target data types are listed in Table 3-1.
Decision Tree is the only algorithm that supports a cost matrix at build time. However, you can create a cost matrix and associate it with any classification model for scoring.
If you want to use costs for scoring, create a table with the columns shown in Table 5-13, and use the DBMS_DATA_MINING.ADD_COST_MATRIX procedure to add the cost matrix table to the model. You can also specify a cost matrix inline when invoking a PREDICTION function.
See Also:
Oracle Data Mining Concepts for information about costs