rfinterval {rfinterval}

Prediction Intervals for Random forests

Description

The rfinterval constructs prediction intervals for random forest predictions using a fast implementation package 'ranger'.

Usage

rfinterval(formula = NULL, train_data = NULL, test_data = NULL,
  method = c("oob", "split-conformal", "quantreg"), alpha = 0.1,
  symmetry = TRUE, seed = NULL, params_ranger = NULL)

Arguments

formula

Object of class formula or character describing the model to fit. Interaction terms supported only for numerical variables.

train_data

Training data of class data.frame, matrix, or dgCMatrix (Matrix).

test_data

Test data of class data.frame, matrix, or dgCMatrix (Matrix).

method

Method for constructing prediction interval. If method = "oob", compute the out-of-bag prediction intervals; if method = "split-conformal", compute the split conformal prediction interval; if method = "quantreg", use quantile regression forest to compute prediction intervals.

alpha

Confidence level. alpha = 0.05 for the 95% prediction interval.

symmetry

True if constructing symmetric out-of-bag prediction intervals, False otherwise. Only for method = "oob"

seed

Seed (only for method = "split-conformal")

params_ranger

List of further parameters that should be passed to ranger. See ranger for possible parameters.

Value

oob_interval

Out-of-bag prediction intervals

sc_interval

Split-conformal prediction intervals

quantreg_interval

Quantile regression forest prediction intervals

alpha

Confidence level for prediction intervals

testPred

Random forest prediction for test set

train_data

Training data

test_data

Test data

References

Haozhe Zhang, Joshua Zimmerman, Dan Nettleton, and Dan Nordman. (2019). "Random Forest Prediction Intervals." The American Statistician. Doi: 10.1080/00031305.2019.1585288.

Haozhe Zhang. (2019). "Topics in Functional Data Analysis and Machine Learning Predictive Inference." Ph.D. Dissertations. Iowa State University Digital Repository. 17929.

Lei, J., Max G’Sell, Alessandro Rinaldo, Ryan J. Tibshirani, and Larry Wasserman. "Distribution-free predictive inference for regression." Journal of the American Statistical Association 113, no. 523 (2018): 1094-1111.

Meinshausen, Nicolai. "Quantile regression forests." Journal of Machine Learning Research 7 (2006): 983-999.

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

Examples

train_data <- sim_data(n = 500, p = 8)
test_data <- sim_data(n = 500, p = 8)
output <- rfinterval(y~., train_data = train_data, test_data = test_data,
                     method = c("oob", "split-conformal", "quantreg"),
                     symmetry = TRUE,alpha = 0.1)
y <- test_data$y
mean(output$oob_interval$lo < y & output$oob_interval$up > y)
mean(output$sc_interval$lo < y & output$sc_interval$up > y)
mean(output$quantreg_interval$lo < y & output$quantreg_interval$up > y)