A Walk-through-guide for Using Decision Trees Software - Data Mining with Decision Trees: Theory and Applications - page 161

Fig. 10.10 The classifier tree is obtained for Iris dataset.
Administration” for instructions how to install a package). Then it trains a
classification tree for the Iris dataset where Species is the target attribute
and all other attributes are used as input attributes. Then in the third line
the plot function is used to visualize c the tree as illustrated in Figure 10.10.
The nodes are numbered. Each leaf node contains a histogram that shows
the probability of an instance to be labeled with each one of the classes.
The tree induction algorithm can be configured using the
ctree_control() function. In particular, the following parameters can
be set:
(1) mincriterion — The threshold value of the splitting criterion that must
be exceeded in order to implement a split.
(2) minsplit — The minimum number of instances in a node in order to be
considered for splitting.
(3) minbucket — The minimum number of instances in a terminal node.
(4) stump — A Boolean parameter that indicates whether a decision stump
should be built.
(5) maxdepth — The maximum depth of the tree.
The following script illustrates a more complicated learning process.
It begins by splitting the Iris data into train (2/3 of the instances) and
test (the remaining instances). Then it trains a classification tree for the
Species target attribute by using only the train data with the following