cross validation

Description

cross validation function for lars algorithm

Usage

1
2
3
HDcvlars(X, y, nbFolds = 10, index = seq(0, 1, by = 0.01),
  mode = c("fraction", "lambda"), maxSteps = 3 * min(dim(X)),
  partition = NULL, intercept = TRUE, eps = .Machine$double.eps^0.5)

Arguments

X

the matrix (of size n*p) of the covariates.

y

a vector of length n with the response.

nbFolds

the number of folds for the cross-validation.

index

Values at which prediction error should be computed. When mode = "fraction", this is the fraction of the saturated |beta|. The default value is seq(0,1,by=0.01). When mode="lambda", this is values of lambda.

mode

Either "fraction" or "lambda". Type of values containing in partition.

maxSteps

Maximal number of steps for lars algorithm.

partition

partition in nbFolds folds of y. Must be a vector of same size than y containing the index of folds.

intercept

If TRUE, there is an intercept in the model.

eps

Tolerance of the algorithm.

Value

A list containing

cv

Mean prediction error for each value of index.

cvError

Standard error of cv.

minCv

Minimal cv criterion.

minIndex

Value of index for which the cv criterion is minimal.

index

Values at which prediction error should be computed. This is the fraction of the saturated |beta|. The default value is seq(0,1,by=0.01).

maxSteps

Maximum number of steps of the lars algorithm.

Author(s)

Quentin Grimonprez

Examples

1
2
dataset=simul(50,10000,0.4,10,50,matrix(c(0.1,0.8,0.02,0.02),nrow=2))
result=HDcvlars(dataset$data,dataset$response,5)