rope: FDR controlled model selection

In rope: Model Selection with FDR Control of Selected Variables

Description

Estimates a model from bootstap counts. The objective is to maximize accuracy while controlling the false discovery rate of selected variables. Developed for high-dimensional models with number of variables in the order of at least 10000.

Usage

rope(data, B, fdr = 0.1, mc.cores = getOption("mc.cores", 2L),
  only.first = FALSE)

Arguments

data	Matrix of variable presence counts. One column for each variable, one row for each parameter value (e.g. levels of regularization).
B	Number of bootstraps used to construct data. At least 21 are needed for u-shape test heuristic to work, but in general it is recommended to use many more.
fdr	Vector of target false discovery rates to return selections for
mc.cores	Number of threads to run in parallel (1 turns of parallelization)
only.first	Skip second part of algorithm. Saves time but gives worse results.

Value

A list with components

selection	matrix (one row for each fdr target, one column for each variable)
q	vector of q-values, one for each variable
level	index of most separating parameter value
alt.prop	estimated proportion of alternative variables

Author(s)

Jonatan Kallus, [e-mail]

Examples

## Not run: 
data # a matrix of selection counts, for 100 bootstraps, with ncol(data)
  # potential variables counted for nrow(data) different penalization levels
fdr <- c(0.05, 0.1)
result <- rope(data, 100, fdr)

## End(Not run)

rope documentation built on May 2, 2019, 12:36 p.m.