fit.lgem: Linear generated effect modifiers method (workhorse)
In syhyunpark/csim: Constrained single index models (CSIM)

Description Usage Arguments Value See Also

lgem is a function for fitting a generated treatment effect modifiers (GEM) model by the "Numerator" criterion of Petkova et al. (2016). The GEM variable is defined as a linear combination of pretreatment covariates X, optimized under a linear model framework to exhibit a strong (linear) interaction effect with the treatment indicator. CSIM is a semiparametric generalization of the GEM model for modeling possilbly nonlinear interaction effects.

1	fit.lgem(y, Tr, X, use.lasso = FALSE, eff.aug = FALSE, n.folds = 20)

`y`	treatment outcomes, a n-by-1 vector
`Tr`	treatment indicators, a n-by-1 vector; each element represents one of the K available treatment options
`X`	pretreatment covarates, a n-by-p matrix
`use.lasso`	if `TRUE`, estimate the treatment-specific linear regression coefficients via the Lasso, with the tuning parameters selected by `n.folds` cross-validation for prediction errors; the default is `FALSE`.
`eff.aug`	if `TRUE`, perform efficiency augmentation (using a L1 regularized linear regression for the main effects of X); the default is `FALSE`.
`n.folds`	an integer value specifying the number of folds to be used in cross-validation.

a list of information of the fitted model including

`alpha.coef`	a vector of the estimated GEM coefficients.
`gamma.list`	a list of the treatment-specific slopes
`coef.list`	a list of the treatment-specific coefficients, i.e., the slope times the GEM coefficients.