Classifier that weights a set of class association rules
That are association rules where exclusively one class attribute-value-pair is allowed in the consequence.No pruning is perfomed (except that the number of mined rules can be limited and from that set the user can choose to consider only the top ranked rules with the -L option).
For classification there are three options:a) all rules are equally weighted with 1 (default)
b) all rules are linearly weighted in a decreasing manner according to the sort order of the mining algorithmc) all rules are weighted with the inverse function 1/rank in sort order.
(based on WEKA 3.7)
For further options, click the 'More' - button in the dialog.
All weka dialogs have a panel where you can specify classifier-specific parameters.
- Type: Data Training data
- Type: Weka 3.7 Classifier Trained model