control_sel constructs a list with all necessary control parameters
for selection model.
control_sel(
est_method = c("mle", "gee"),
gee_h_fun = 1,
optimizer = c("maxLik", "optim"),
maxlik_method = c("NR", "BFGS", "NM"),
optim_method = c("BFGS", "Nelder-Mead"),
epsilon = 1e-04,
maxit = 500,
trace = FALSE,
penalty = c("SCAD", "lasso", "MCP"),
a_SCAD = 3.7,
a_MCP = 3,
lambda = -1,
lambda_min = 0.001,
nlambda = 50,
nfolds = 10,
print_level = 0,
start_type = c("zero", "mle", "naive"),
nleqslv_method = c("Broyden", "Newton"),
nleqslv_global = c("dbldog", "pwldog", "cline", "qline", "gline", "hook", "none"),
nleqslv_xscalm = c("fixed", "auto"),
dependence = FALSE,
key = NULL
)List with selected parameters.
Method of estimation for propensity score model ("mle" or "gee"; default is "mle").
Smooth function for the generalized estimating equations (GEE) method.
(for the est_method="mle" only) optimization function for maximum likelihood estimation.
(for the est_method="mle" only) maximisation method that will be passed to maxLik::maxLik() function. Default is NR.
(for the est_method="mle" only) maximisation method that will be passed to stats::optim() function. Default is BFGS.
Tolerance for fitting algorithms by default 1e-6.
Maximum number of iterations.
logical value. If TRUE trace steps of the fitting algorithms. Default is FALSE
The penalization function used during variables selection.
The tuning parameter of the SCAD penalty for selection model. Default is 3.7.
The tuning parameter of the MCP penalty for selection model. Default is 3.
A user-specified \(\lambda\) value during variable selection model fitting.
The smallest value for lambda, as a fraction of lambda.max. Default is .001.
The number of lambda values. Default is 50.
The number of folds for cross validation. Default is 10.
this argument determines the level of printing which is done during the optimization (for propensity score model) process.
Type of method for start points for model fitting taking the following values
if zero then start is a vector of zeros (default for all methods).
if mle (for est_method="gee" only) starting parameters are taken from the result of the est_method="mle" method.
(for the est_method="gee" only) The method that will be passed to nleqslv::nleqslv() function.
(for the est_method="gee" only) The global strategy that will be passed to nleqslv::nleqslv() function.
(for the est_method="gee" only) The type of x scaling that will be passed to nleqslv::nleqslv() function.
logical value (default TRUE) informing whether samples overlap (NOT YET IMPLEMENTED, FOR FUTURE DEVELOPMENT).
binary key variable allowing to identify the overlap (NOT YET IMPLEMENTED, FOR FUTURE DEVELOPMENT).
Smooth function (gee_h_fun) for the generalized estimating equations (GEE) method taking the following values
if 1 then \(\boldsymbol{h}\left(\boldsymbol{x}, \boldsymbol{\theta}\right) =
\frac{\pi(\boldsymbol{x}, \boldsymbol{\theta})}{\boldsymbol{x}}\),
if 2 then \( \boldsymbol{h}\left(\boldsymbol{x}, \boldsymbol{\theta}\right) = \boldsymbol{x}\)
nonprob() -- for fitting procedure with non-probability samples.