stabsel {mboost}R Documentation

Stability Selection

Description

Selection of influential variables or model components with error control.

Usage

stabsel(object, FWER = 0.05, cutoff, q, 
        folds = cv(model.weights(object), type = "subsampling"),
        papply = if (require("multicore")) mclapply else lapply, ...)

Arguments

object

an mboost object.

FWER

family-wise error rate to be controlled by the selection procedure.

cutoff

cutoff between 0.5 and 1.

q

average number of selected variables.

folds

a weight matrix with number of rows equal to the number of observations, see cvrisk.

papply

(parallel) apply function. In the absence of package multicore sequential computations via lapply are performed. Alternatively, parallel computing via mclapply (package multicore), or clusterApplyLB (package snow) can be used. In the latter case, usually more setup is needed (see example for some details).

...

additional arguments to cvrisk.

Details

This function implements the "stability selection" procedure by Meinshausen and Buehlmann (2010).

Either cutoff or q must be specified. The probability of selecting at least one non-influential variable (or model component) is less than FWER.

Value

An object of class stabsel with elements

phat

selection probabilities.

selected

elements with maximal selection probability greater cutoff.

max

maximum of selection probabilities.

cutoff

cutoff used.

q

average number of selected variables used.

FWER

family-wise error rate.

References

N. Meinshausen and P. Buehlmann (2010), Stability selection. Journal of the Royal Statistical Society, Series B, 72(4).

Examples



  ### (too) low-dimensional example    
  sbody <- stabsel(glmboost(DEXfat ~ ., data = bodyfat), q = 3,
                   papply = lapply)
  sbody
  opar <- par(mai = par("mai") * c(1, 1, 1, 2.7))
  plot(sbody)
  par(opar)


[Package mboost version 2.1-1 Index]