Databases Reference
In-Depth Information
two other classification modelers: support vector machine (SVM) and artificial
neural network (ANN).
Close the confusion matrix for the decision tree classifier.
Drag the SVM classifier (support vector machine) to the Iris dataset.
After processing is complete, hover over each of the two models to view
and compare error rates.
At this point, we shift focus to ANN model construction. We will get back to
the decision tree and SVM models once the ANN has been built.
The VisMiner ANN classifier implements a feature that allows the user to
interact with the algorithm during the model building process. For this reason,
its execution is a little more complex than that of the simpler drag and drop of
the decision tree and SVM modelers.
Drag the ANN classifier to the Iris dataset.
Select “Build interactively”.
The interactive build option supports user control over the training process
(Figure 5.2). The effectiveness of ANN training depends on the network
learning rate and momentum, yet there is no single best learning rate or
momentum for all datasets. It will vary from dataset to dataset. The interactive
build option allows you to monitor the training progress while adjusting the
learning rate and momentum as it progresses.
Figure 5.2
ANN Interactive Training
