The foehnix function allows to estimate
a regularized logistic regression model for the concomitants.
In case glmnet.control is provided the
EM algorithm uses this function to estimate the regression coefficients
of the concomitant model. glmnet.control allows
to specify options forwarded to glmnet::glmnet except
family ("binomial") and intercept (TRUE).
Depending on glmnet.control the AIC, BIC, or
log-likelihood criterion will be used.
foehnix_glmnet(y, x, arg)response vector (binary)
design matrix. If a column "(Intercept)" exists it will be
dropped (as we force glmnet to estimate an intercept).
list of arguments forwarded to glmnet::glmnet.