Calculate mu(Xk) where Xk denotes the null sample

This function calculates mu(Xk) for null replicates Xk.

calulate.mu_Xk(
  binary = FALSE,
  X,
  i2,
  S,
  out,
  nulls.list_S = NULL,
  gamma_X.list_S = NULL,
  useMC = TRUE,
  Xmodel = "gaussian",
  algo = "lasso",
  predict.fun,
  cv.rule = "min",
  verbose = FALSE
)

Arguments

binary	whether the response variable is binary (will be converted to factors if TRUE; default: FALSE).
X	a n by p matrix, containing all the covariates.
i2	the index of inference samples.
S	a list of selected variables.
out	the fitted model from train.fun.
nulls.list_S	a list of length |S| whose element is a (|i2|*K)-dimensional vector, which contains K set of null samples.
gamma_X.list_S	a list of length |S|, with each element being the linear coefficient of the given covariate on the other covariates (only relevant when Xmodel = "gaussian"; default: NULL).
useMC	whether to use Monte Carlo estimators of the conditional quantities (default: TRUE).
Xmodel	model of the covaraites (default: "gaussian").
algo	a fitting algorithm (default: "lasso").
predict.fun	a function to produce predictions of the response variable with a given fitted model from train.fun and a matrix of new covariate values.
cv.rule	indicates which rule should be used for the predict function, either "min" (the usual rule) or "1se" (the one-standard- error rule); default: "min"). See the glmnet help files for details.
verbose	whether to show intermediate progress (default: FALSE).

Value

A list of kength |S|, whose element is the matrix of mu(Xk) with dimension n2-by-K or n2-by-1.

References

Zhang L, Janson L (2020). “Floodgate: Inference for Model-Free Variable Importance.” arXiv preprint arXiv:2007.01283.

See also

Other methods: calculate.V_mean(), fg.inference.binary(), fg.inference(), fit.mu(), floodgate.binary(), floodgate(), inference_general()