simulation.alternative: The alternative of running simulation
In fadedsoul/cdml: CDML

The alternative of running simulation

simulation.alternative(model = "CTE", rep = 7, file = "simu.alter.rds",
  method = c("CDML"), simu.times = 2, g.method = c("rf", "nnet"),
  gps.method = c("linear&normal"), fold = c(1),
  sample.size.option = c(200, 300, 500, 800, 1000, 1500, 2000),
  num.cov.option = c(1, 2, 3, 5, 8, 10, 15), treatment.sd.option = c(1, 2,
  3, 5, 8, 10), response.sd.option = c(1, 2, 3, 5, 8, 10),
  struct.model.option = c("linear", "polynom", "polynom2", "polynom3"),
  detoured = FALSE, keep.same = TRUE)

`model`	model varies from "IV", "CTE"
`rep`	within each set-up, number of repetition that will be done
`file`	file that we save our return into
`method`	estimation method we choose from double machine learning method ("CDML") and simple regression method ("SR"), Hirano & Imbens method ("HI")
`simu.times`	number of simulations one would run
`g.method`	a vector of method for regression estimation
`gps.method`	a vector of method for generalized propensity score estimation
`fold`	number of folds for sample splitting
`sample.size.option`	a vector of number of samples that will be selected from
`num.cov.option`	a vector of number of covariates that will be selected from
`treatment.sd.option`	a vector of standard deviation of treatment generation that will be selected from
`response.sd.option`	a vector of standard deviation of response generation that will be selected from, assuming normal distributed
`struct.model.option`	a vector of parametric curves of response generation, typically choose from "linear", "polynom", "polynom2", "polynom3", "mixture".
`detoured`	if FALSE method will be done normally; otherwise a special simulation will be examined, where CDML is chosen method, CTE is chosen model, gps.method "series" is not allowed.
`keep.same`	in the case detoured is true, whether we use the same gps.method to estimate T based on X