stability function

Function to perform stability selection

stability(x, y, EV, threshold = 0.75, B = 100, fraction = 0.5,
          model.selector = lasso.firstq, args.model.selector = NULL,
          parallel = FALSE, ncores = getOption("mc.cores", 2L),
          verbose = FALSE)

Arguments

• x: Design matrix (without intercept).
• y: Response vector.
• EV: Bound for expected number of false positives.
• threshold: Threshold for selection frequency. Must be in (0.5, 1).
• B: Number of sub-sample iterations.
• fraction: Fraction of data used at each of the B sub-samples.
• model.selector: Function to perform model selection. Default is lasso.firstq. User supplied function must have at least three arguments: x (the design matrix), y (the response vector) and q (the maximal model size). Return value is the index vector of selected columns. See lasso.firstq for an example. Additional arguments can be passed through args.model.selector.
• args.model.selector: Named list of further arguments for function model.selector.
• parallel: Should parallelization be used? (logical)
• ncores: Number of cores used for parallelization.
• verbose: Should information be printed out while computing (logical).

Returns

• selected: Vector of selected predictors.

• freq: Vector of selection frequencies.

• q: Size of fitted models in order to control error rate at desired level.

References

Meinshausen, N. and , P. (2010) Stability selection (with discussion). Journal of the Royal Statistical Society: Series B 72 , 417--473.

, P., Kalisch, M. and Meier, L. (2014) High-dimensional statistics with a view towards applications in biology. Annual Review of Statistics and its Applications 1 , 255--278

Author(s)

Lukas Meier

Examples

x <- matrix(rnorm(100 * 1000), nrow = 100, ncol = 1000)
y <- x[,1] * 2 + x[,2] * 2.5 + rnorm(100)
fit.stab <- stability(x, y, EV = 1)
fit.stab
fit.stab$freq[1:10] ## selection frequency of the first 10 predictors