treemisc (version 0.0.1)

gen_mease: Generate data from the Mease model

Description

Generate binary classification data from the Mease model Mease et al. (2007).

Usage

gen_mease(n = 1000, nsim = 1)

Value

A data frame with 3 + nsim columns. The first two columns give the values of the numeric features x1 and x2. The third column (yprob) gives the true probabilities (i.e., PrY = 1 | X = x). The remaining nsim columns (yclass<i>, i = 1, 2, ..., nsim) give the simulated binary outcomes corresponding to yprob.

Arguments

n

Integer specifying the number of observations. Default is 1000.

nsim

Integer specifying the number of binary repsonses to generate. Default is 1.

References

Mease D, Wyner AJ, Buja A. Boosted classification trees and class probability quantile estimation. Journal of Machine Learning Research. 2007; 8:409–439.

Examples

# Generate N = 1000 observations from the Mease model
set.seed(2254)  # for reproducibility 
mease <- gen_mease(1000, nsim = 1)

# Plot predictor values colored by binary outcome
cols <- palette.colors(2, palette = "Okabe-Ito", alpha = 0.3)
plot(x2 ~ x1, data = mease, col = cols[mease$yclass1 + 1], pch = 19)