cv.gglasso {gglasso}

Cross-validation for gglasso

Description

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

Usage

cv.gglasso(
  x,
  y,
  group,
  lambda = NULL,
  pred.loss = c("misclass", "loss", "L1", "L2"),
  nfolds = 5,
  foldid,
  delta,
  ...
)

Arguments

x

matrix of predictors, of dimension n \times p; each row is an observation vector.

y

response variable. This argument should be quantitative for regression (least squares), and a two-level factor for classification (logistic model, huberized SVM, squared SVM).

group

a vector of consecutive integers describing the grouping of the coefficients (see example below).

lambda

optional user-supplied lambda sequence; default is NULL, and gglasso chooses its own sequence.

pred.loss

loss to use for cross-validation error. Valid options are:

• "loss" for classification, margin based loss function.

• "misclass" for classification, it gives misclassification error.

• "L1" for regression, mean square error used by least squares regression loss="ls", it measure the deviation from the fitted mean to the response.

• "L2" for regression, mean absolute error used by least squares regression loss="ls", it measure the deviation from the fitted mean to the response.

Default is "loss".

nfolds

number of folds - default 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 allowable is nfolds=3.

foldid

an optional vector of values between 1 and nfold identifying what fold each observation is in. If supplied, nfold can be missing.

delta

parameter \delta only used in huberized SVM for computing log-likelihood on validation set, only available with pred.loss = "loss", loss = "hsvm".

...

other arguments that can be passed to gglasso.

Details

The function runs gglasso nfolds+1 times; the first to get the lambda sequence, and then the remainder to compute the fit with each of the folds omitted. The average error and standard deviation over the folds are computed.

Value

an object of class cv.gglasso 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.

name

a text string indicating type of measure (for plotting purposes).

gglasso.fit

a fitted gglasso 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)

Yi Yang and Hui Zou
Maintainer: Yi Yang <[e-mail]>

References

Yang, Y. and Zou, H. (2015), “A Fast Unified Algorithm for Computing Group-Lasso Penalized Learning Problems,” Statistics and Computing. 25(6), 1129-1141.
BugReport: https://github.com/emeryyi/gglasso

See Also

gglasso, plot.cv.gglasso, predict.cv.gglasso, and coef.cv.gglasso methods.

Examples

# load gglasso library
library(gglasso)

# load data set
data(bardet)

# define group index
group <- rep(1:20,each=5)

# 5-fold cross validation using group lasso 
# penalized logisitic regression
cv <- cv.gglasso(x=bardet$x, y=bardet$y, group=group, loss="ls",
pred.loss="L2", lambda.factor=0.05, nfolds=5)