party (version 0.9-10)

cforest: Random Forest

Description

An implementation of the random forest and bagging ensemble algorithms utilizing conditional inference trees as base learners.

Usage

cforest(formula, data = list(), subset = NULL, weights = NULL, 
        controls = cforest_control(),
        xtrafo = ptrafo, ytrafo = ptrafo, scores = NULL)
varimp(x, mincriterion = 0.0)

Arguments

formula

a symbolic description of the model to be fit.

data

an data frame containing the variables in the model.

subset

an optional vector specifying a subset of observations to be used in the fitting process.

weights

an optional vector of weights to be used in the fitting process. Only non-negative integer valued weights are allowed.

controls

an object of class ForestControl-class, which can be obtained using cforest_control.

xtrafo

a function to be applied to all input variables. By default, the ptrafo function is applied.

ytrafo

a function to be applied to all response variables. By default, the ptrafo function is applied.

scores

an optional named list of scores to be attached to ordered factors.

x

an object as returned by cforest.

mincriterion

the value of the test statistic or 1 - p-value that must be exceeded in order make use of a split. See ctree_control.

Value

• An object of class RandomForest-class.

Details

This implementation of the random forest (and bagging) algorithm differs from the reference implementation in randomForest with respect to the base learner used and the aggregation scheme applied. Conditional inference trees, see ctree, are fitted to each of the ntree (defined via cforest_control) bootstrap samples of the learning sample. There are many hyper parameters that can be controlled, see cforest_control. You MUST NOT change anything you don't understand completely.

The aggregation scheme works by averaging observation weights extracted from each of the ntree trees and NOT by averaging predictions directly. See Hothorn et al. (2004) for a description.

Ensembles of conditional inference trees have not yet been extensively tested, so this routine is meant for the expert user only and its current state is rather experimental. However, there are some things that can't be done with randomForest, for example fitting forests to censored response variables or to multivariate and ordered responses.

By default, raw test statitics are maximized and five inputs are randomly examined for possible splits in each node. Note that this implies biased internal variable selection which might affect variable importance measures derived from such a forest.

Function varimp can be used to compute variable importance measures similar to those computed by importance.

References

Leo Breiman (2001). Random Forests. Machine Learning, 45(1), 5--32.

Torsten Hothorn, Berthold Lausen, Axel Benner and Martin Radespiel-Troeger (2004). Bagging Survival Trees. Statistics in Medicine, 23(1), 77--91.

Torsten Hothorn, Peter Buhlmann, Sandrine Dudoit, Annette Molinaro and Mark J. van der Laan (2006). Survival Ensembles. Biostatistics, 7(3), 355--373.

Examples

### honest (i.e., out-of-bag) cross-classification of
    ### true vs. predicted classes
    table(mammoexp$ME, predict(cforest(ME ~ ., data = mammoexp, 
                               control = cforest_control(ntree = 50)),
                               OOB = TRUE))

    ### fit forest to censored response
    if (require("ipred")) {

        data("GBSG2", package = "ipred")
        bst <- cforest(Surv(time, cens) ~ ., data = GBSG2, 
                   control = cforest_control(ntree = 50))

        ### estimate conditional Kaplan-Meier curves
        treeresponse(bst, newdata = GBSG2[1:2,], OOB = TRUE)

        ### if you can't resist to look at individual trees ...
        party:::prettytree(bst@ensemble[[1]], names(bst@data@get("input")))
    }

Run the code above in your browser using DataCamp Workspace