tune.surv.interval: Wrapper function for conditional logistic lasso objects.

In zhizuio/mixlasso: Mix lasso models

Wrapper function for conditional logistic lasso objects.

Description

Wrapper function for conditional logistic lasso objects optized based on the full likelihood by the EPSGO algorithm. This function is mainly used within the function epsgo.

Usage

tune.surv.interval(
  parms,
  x = x,
  y = y,
  x_test = NULL,
  y_test = NULL,
  family = "cox",
  strata.surv = NULL,
  intercept = TRUE,
  alpha = 0,
  lambda = NULL,
  num.nonpen = 0,
  nfolds = 3,
  foldid = NULL,
  seed = 12345,
  standardize.response = FALSE,
  p = NULL,
  type.measure = "deviance",
  type.min = "lambda.min",
  parallel = FALSE,
  verbose = FALSE,
  search.path = FALSE,
  ...
)

Arguments

x, y	input matrix.
family	response type.
strata.surv	stratification variable for the Cox survival model.
intercept	should intercept(s) be fitted (default=TRUE) or set to zero (FALSE).
lambda	optional user-supplied lambda sequence; default is NULL, and espsgo chooses its own sequence.
foldid	an vector of values for the cross-validation.
seed	random seed
standardize.response	standardization for the response variables. Default: TRUE.
p	the number of predictors from different data source.
parallel	If TRUE, use parallel foreach to fit each fold except parallelizing each lambda for the tree-lasso methods. If c(TRUE,TRUE), use parallel foreach to fit each fold and each lambda.
verbose	print the middle search information, default is TRUE.

Details

mixlasso

Value

An object with list "tune.clogit.interval"

q.val	the minimum MSE (or minus likelihood for the Cox model) through the cross-validation
model	some model related quantities: alpha the optimzed alpha lambda the optimzed (first) lambda ipf the optimzed penalty factors p a vector of the numbers of features from multiple data sources nfolds number of folds used for the cross-validation cvreg the cross-validation results

zhizuio/mixlasso documentation built on March 21, 2022, 1:07 a.m.