Database Reference
In-Depth Information
7) In Figure 13-5, add the Decision Tree operator in the Training side of the cross-validation
sub-process, and the Apply Model operator on the Testing side. Leave the Decision Tree's
operator as gain_ratio for now. The splines depicted here are automatically drawn when
you drag these operators into these two areas. If for any reason you do not have these
splines configured in this way, connect the ports as shown so that your sub-process
matches Figure 13-5. We must now complete the Testing portion of the sub-process. In
the Operators search field, search for an operator called 'Performance'. There are a
number of these. We will use the first onw: Performance (Classification). The reason for
this is that a decision tree predicts a classification in an attribute—in our example, the
adopter class (innovator, early adopter, etc.).
Figure 13-6. The configuration of the cross-validation sub-process.
8) Once your sub-process is configured, click the blue up arrow to return to the main process.
Connect the mod, tra and ave ports to res ports as shown in Figure 13-7. The mod port will
generate the visual depiction of our decision tree, the tra port will create the training data
set's attribute table, and the avg port will calculate a True Positive table showing the training
data set's ability to predict accurately.
Search WWH ::




Custom Search