bigctree-class: Classification Trees in Random Forests

In gboris/bigrf: Big Random Forests: Classification and Regression Forests for Large Data Sets

Description

Class representing a tree in a classification random forest.

Objects from the Class

Objects can be created by calls of the form new("bigctree", ...), but most often are generated when building a bigcforest using bigrfc.

Slots

insamp:

Object of class "integer". Number of times each training example is in the bag for this tree.

inweight:

Object of class "numeric". Sum of class weights for each in-bag example.

nnodes:

Object of class "integer". Number of nodes in the tree.

treemap:

Object of class "matrix". Indices of tree nodes. For a given node n, treemap[n, ] represents the indices of the left and right children. If a node is terminal, then both elements of treemap[n, ] are 0.

nodeclass:

Object of class "integer". Majority class of each terminal node.

nodewt:

Object of class "numeric". Sum of class weights of the examples in each terminal node.

bestvar:

Object of class "integer". Best variable that each node was split on, or 0 for terminal nodes.

bestnumsplit:

Object of class "numeric". Split point for nodes split on a numerical variable.

bestcatsplit:

Object of class "list". Split point, expressed as integer vectors indicating which factor levels to split on, for nodes split on a categorical variable.

termincount:

Object of class "numeric". Number of training examples in each terminal node.

trainprednode:

Object of class "integer". Terminal node number for each training example.

trainpredclass:

Object of class "integer". Class of terminal node for each training example.

tgini:

Object of class "numeric". Total gini.

Methods

No methods defined with class "bigctree" in the signature.

gboris/bigrf documentation built on May 16, 2019, 10:14 p.m.