| e) Group the values of your attributes | g) Induce a tree from the data | |
The MinedTree object can use many attributes (i.e. fields in your data source table) to build a tree. However, in the interests of speed you should consider excluding some of the attributes. To aid you in this process the Rank Fields option will rank the attributes in accordance with their effect on the outcome. Attributes at the lower end of the ranking order should then be considered for exclusion - and their usage set to Excluded via the Field Usage tab.
Excluding data fields from being used is especially beneficial when such data fields have large numbers of string values. For example, consider a customer database with a unique index identifier or account ID field. Such a field will have no predictive value for the purposes of tree induction data mining. A data field holding unique values could not help classify the data.