Earth Learner

Channels

Inputs:

• Data (Table)

Outputs:

• Learner

  The Earth learning algorithm with parameters as specified in the dialog.

• Predictor

  Trained regressor

• Basis Matrix (Table)

  A data table of induced basis terms.

Signal Predictor sends the regressor only if signal Data is present.

Description

This widget constructs a Earth learning algorithm (an implementation of the MARS - Multivariate Adaptive Regression Splines). As all widgets for classification and regression, this widget provides a learner and classifier/regressor on the output. Learner is a learning algorithm with settings as specified by the user. It can be fed into widgets for testing learners, for instance Test Learners.

1. Learner/Predictor can be given a name under which it will appear in other widgets (say Test Learners or Predictions).
2. The Max. term degree parameter specifies the degree of the terms induced in the forward pass. For instance, if set to 1 the resulting model will contain only linear terms.
3. The Max. terms specifies how many terms can be induces in the forward pass. A special value Automatic instructs the learner to set the limit automatically based on the dimensionality of the data (min(200, max(20, 2 * NumberOfAttributes)) + 1)
4. The Knot penalty is used in the pruning pass (hinge function penalty for the GCV calculation)

After changing one or more settings, you need to push 5 Apply, which will put the new learner on the output and, if the training examples are given, construct a new predictor and output it as well.

Examples

Lets use the learner to train a model on a data subset and test it on unseen instances.