ccboost (version 0.1-1.3)

dataLS: generate random data for classification as in Long and Servedio (2010)

Description

generate random data for classification as in Long and Servedio (2010)

Usage

dataLS(ntr, ntu = ntr, nte, percon)

Arguments

ntr

number of training data

ntu

number of tuning data, default is the same as ntr

nte

number of test data

percon

proportion of contamination, must between 0 and 1. If percon > 0, the labels of the corresponding percenrage of response variable in the training and tuning data are flipped.

Value

a list with elements xtr, xtu, xte, ytr, ytu, yte for predictors of disjoint training, tuning and test data, and response variable -1/1 of training, tuning and test data.

References

P. Long and R. Servedio (2010), Random classification noise defeats all convex potential boosters, Machine Learning Journal, 78(3), 287--304.

Examples

# NOT RUN {
dat <- dataLS(ntr=100, nte=100, percon=0)
# }