Data Mining: Practical Machine Learning Tools and ... - LIDeCC

13.2 THE STRUCTURE OF WEKA 455Figure 13.2 (continued)The presence of a main() method in a class indicates that it can be run from thecommand line and that all learning methods and filter algorithms implementit.Other packagesSeveral other packages listed in Figure 13.1(a) are worth mentioning: weka.associations,weka.clusterers, weka.estimators, weka.filters, and weka.attributeSelection.The weka.associations package contains association rule learners. Thesehave been placed in a separate package because association rules are fundamentallydifferent from classifiers. The weka.clusterers package containsmethods for unsupervised learning. The weka.estimators package contains subclassesof a generic Estimator class, which computes different types of probabilitydistribution. These subclasses are used by the Naïve Bayes algorithm(among others).In the weka.filters package, the Filter class defines the general structure ofclasses containing filter algorithms, which are all implemented as subclasses ofFilter. Like classifiers, filters can be used from the command line: we will seehow shortly. The weka.attributeSelection package contains several classesfor attribute selection. These are used by the AttributeSelectionFilter inweka.filters.supervised.attribute, but can also be invoked separately.