predict.forest: Predict method for random decision forest classifier fits

In bert9bert/ParallelForest: Random Forest Classification with Parallel Computing

Description

Predict method for random decision forest classifier fits

Usage

## S4 method for signature 'forest'
predict(object, newdata, type, decision_threshold, ...)

Arguments

object	Object of class inheriting from "forest-class".
newdata	A data frame in which to look for variables with which to predict.
type	Set to "response" to return predictions as probabilities. Set to "binary_0_1" to apply the decision rule threshold to return predictions as 0s and 1s (default).
decision_threshold	If type="binary_0_1", then predicted probabilities above this threshold will be returned as 1, otherwise returned as 0. Default is 0.5 (majority rules).
...	further arguments passed to or from other methods.

Examples

  data(easy_2var_data)

  fforest = grow.forest(Y~X1+X2, data=easy_2var_data,
    min_node_obs=5, max_depth=10,
    numsamps=90, numvars=1, numboots=5)

  xnew = data.frame(
    X1 = c(0.06, 0.05, 0.05, 0.01, 0.09, 0.05, 0.05,-1000, 1000),
    X2 = c(0.03, 0.02, 0.05, 0.03, 0.04,-1000, 1000, 0.04, 0.03)
    )

  fforest_ynewhat = predict(fforest, xnew)

bert9bert/ParallelForest documentation built on Feb. 29, 2020, 2:34 p.m.