cv.cpernet: Cross-validation for cpernet

In SALES: The (Adaptive) Elastic Net and Lasso Penalized Sparse Asymmetric Least Squares (SALES) and Coupled Sparse Asymmetric Least Squares (COSALES) using Coordinate Descent and Proximal Gradient Algorithms

Cross-validation for cpernet

Description

Does k-fold cross-validation for cpernet, produces a plot, and returns a value for lambda. This function is based on the cv function from the glmnet package.

Usage

cv.cpernet(
  x,
  y,
  w = 1,
  lambda = NULL,
  pred.loss = "loss",
  nfolds = 5,
  foldid,
  tau = 0.8,
  ...
)

Arguments

x	x matrix as in cpernet.
y	response variable y as in cpernet.
w	weight applied to the asymmetric squared error loss of the mean part. Default is 1.0.
lambda	optional user-supplied lambda sequence; default is NULL, and cpernet chooses its own sequence.
pred.loss	loss function used to calculate cross-validation error. The only option now is "loss", which is the asymmetric squared error loss (ASEL).
nfolds	number of folds. Default value is 5. Although nfolds can be as large as the sample size (leave-one-out CV), it is not recommended for large datasets. Smallest value allowed is 3.
foldid	an optional vector of values between 1 and nfolds, identifying what fold each observation is in. If supplied, nfolds will be supressed.
tau	the asymmetry coefficient τ used in the asymmetric squared error loss.
...	other arguments that can be passed to cpernet.

Details

The function runs cpernet nfolds+1 times. The first gets the lambda sequence, and the remainder fits the model with each of the folds removed. The average error and standard deviation over the folds are computed.

Value

an object of class cv.cpernet is returned, which is a list with the ingredients of the cross-validation fit.

lambda	the values of lambda used in the fits.
cvm	the mean cross-validated error - a vector of length length(lambda).
cvsd	estimate of standard error of cvm.
cvupper	upper curve = cvm+cvsd.
cvlower	lower curve = cvm-cvsd.
nzero	a list of two components, each representing the number of non-zero coefficients at each lambda in the mean and scale part.
name	a text string indicating type of measure (for plotting purposes).
cpernet.fit	a fitted cpernet object for the full data.
lambda.min	The optimal value of lambda that gives minimum cross validation error cvm.
lambda.1se	The largest value of lambda such that error is within 1 standard error of the minimum.

Author(s)

Yuwen Gu and Hui Zou

Maintainer: Yuwen Gu <[e-mail]>

See Also

cpernet

Examples

set.seed(1)
n <- 100
p <- 400
x <- matrix(rnorm(n * p), n, p)
y <- rnorm(n)
tau <- 0.30
pf <- abs(rnorm(p))
pf2 <- abs(rnorm(p))
w <- 2.0
lambda2 <- 1
m2.cv <- cv.cpernet(y = y, x = x, w = w, tau = tau, eps = 1e-8,
                    pf.mean = pf, pf.scale = pf2,
                    standardize = FALSE, lambda2 = lambda2)

SALES documentation built on Aug. 16, 2022, 1:05 a.m.