Package

weka.classifiers.trees

Synopsis

Class for building a best-first decision tree classifier. This class uses binary split for both nominal and numeric attributes. For missing values, the method of 'fractional' instances is used.

For more information, see:

Haijian Shi (2007). Best-first decision tree learning. Hamilton, NZ.

Jerome Friedman, Trevor Hastie, Robert Tibshirani (2000). Additive logistic regression : A statistical view of boosting. Annals of statistics. 28(2):337-407.

Available in Weka 3.6.x - 3.7.1. Available via the package management system for Weka >= 3.7.2 (bestFirstTree).

Options

The table below describes the options available for BFTree.

Option	Description
debug	If set to true, classifier may output additional info to the console.
heuristic	If heuristic search is used for binary split for nominal attributes.
minNumObj	Set minimal number of instances at the terminal nodes.
numFoldsPruning	Number of folds in internal cross-validation.
pruningStrategy	Sets the pruning strategy.
seed	The random number seed to be used.
sizePer	The percentage of the training set size (0-1, 0 not included).
useErrorRate	If error rate is used as error estimate. if not, root mean squared error is used.
useGini	If true the Gini index is used for splitting criterion, otherwise the information is used.
useOneSE	Use the 1SE rule to make pruning decision.

Capabilities

The table below describes the capabilites of BFTree.

Capability	Supported
Class	Binary class, Nominal class
Attributes	Unary attributes, Binary attributes, Empty nominal attributes, Nominal attributes, Missing values, Numeric attributes
Min # of instances	1