cml_lasso_win_lbfgs_optimized_parLapply: Title Fitting cmlDIF with L1-penalty via lbfgs (opitmized...

In submission-user/cmlDIFlasso: DIF-detection with Lasso-methods in a cML-Framework

Description

Title Fitting cmlDIF with L1-penalty via lbfgs (opitmized with lambda_max)

Usage

cml_lasso_win_lbfgs_optimized_parLapply(
  data,
  covariates,
  DIF_type = "all.interactions",
  nlambdas = 21,
  tolerance_df = 0,
  no_of_cores_to_use = ifelse(detectCores() == 1, 1, detectCores() - 1),
  ...
)

Arguments

data	matrix of responses
covariates	Matrix of covariate values of all persons (n x ncovs)
DIF_type	can be "items", "variables", "all.interactions", "none"
nlambdas	number of lambda parameters evenly spread between 10^-5 and lambda_max, which is calculated automatically
tolerance_df	Value under which parameters are considered equal to zero for the calculation of degreees of freedom
no_of_cores	Number of cores to be used for parallel computing

Value

list of lbfgs-optimization results for each lambda, the picked model and its parameters, the computing time

submission-user/cmlDIFlasso documentation built on Dec. 23, 2021, 6:40 a.m.