AdaBoost

An ensemble meta-algorithm that combines multiple weak learners to build to build more accurate prediction models.

Signals

Inputs:

• Data

  A data set.

• Preprocessor

  Preprocessed data.

• Learner

  A learning algorithm.

Outputs:

• Learner

  AdaBoost learning algorithm with settings as specified in the dialog.

• Classifier

  Trained classifier (a subtype of Classifier). The AdaBoost classifier signal sends data only if the learning data (signal Data) is present.

Description

The AdaBoost (short for “Adaptive boosting”) widget is a machine-learning algorithm, formulated by Yoav Freund and Robert Schapire. It can be used with other learning algorithms to boost their performance. It does so by tweaking the weak learners.

1. The learner can be given a name under which it will appear in other widgets. The default name is “AdaBoost”.
2. Set the parameters. The base estimator is a tree and you can set:
   • the Number of estimators
   • the Learning rate: it determines to what extent the newly acquired information will override the old information (0 = the agent will not learn anything, 1 = the agent considers only the most recent information)
   • the Algorithm: SAMME (updates base estimator’s weights with classification results) or SAMME.R. (updates base estimator’s weight with probability estimates)
3. Produce a report.
4. Click Apply after changing the settings. That will put the new learner in the output and, if the training examples are given, construct a new classifier and output it as well. To communicate changes automatically tick Apply Automatically.

Examples

For our first example, we loaded the Iris data set and compared the results of two different classification algorithms against the AdaBoost widget.

For our second example, we loaded the Iris data set, sent the data instances to several different classifiers (AdaBoost, Classification Tree, Logistic Regression) and output them in the Predictions widget.