| Parameter Name |
Type |
Range |
Description |
| COMPLEXITY_PENALTY |
Float |
0–1 |
Used to restrict the growth of the tree. The decision tree algorithm produces splits that result in the creation of multiple branches. A low value for COMPEXITY_PENALTY indicates that splits will occur more frequently, and a high value indicates that splits will occur less frequently. The default value varies depending on the number of attributes specified for the mining model. |
| MINIMUM_LEAF_CASES |
Integer |
0–2,147,483,647 |
Similar to the COMPLEXITY_PENALTY in that it controls the number of splits that occur. This parameter, though, is based on the number of leaf nodes present in the tree; a low value indicates that splits will occur more frequently, and a high value indicates that they will occur less frequently. The default value is 10. |
Table 2
Hold the Number of Splits in Check.
The decision trees algorithm contains a couple mining-model parameters for holding down the number of splits. You set these parameters in the Additional Parameters property using the Relational Mining Model Editor.
|