tune_mtry: Tune mtry in final random forest.

Description Usage Arguments Value Note

View source: R/tuning_functions.R

Description

Calculates mse over different values of mtry, returns a plot displaying how the mse changes, and returns the value of mtry with minimum mse.

Usage

1
tune_mtry(object, X, y, step_size = 1, mtry_grid = NULL, main = NULL)

Arguments

object

An object of type fuzzy_forest.

X

A data.frame containing the features selected by fuzzy forests. The names of selected features must be among the column names of X. If X is the training set, the resulting mse's may be biased.

y

The corresponding set of outcomes.

step_size

By default, mtry starts out high (equal to the number of features selected) and goes down by increments of size step_size. selected by fuzzy forests.

mtry_grid

A grid of values of mtry to test. If mtry_grid is non-null, the mse will be calculated for each specified value. In this case, the argument of step_size will be ignored.

main

Title of resulting plot.

Value

An object of type tune_mtry.

Note

This work was partially funded by NSF IIS 1251151.


OHDSI/FuzzyForest documentation built on May 9, 2017, 3:26 p.m.