EAM: Main function to run the EAM algorithm

Description

This function implements the EAM search strategy as described in Kaido, Molinari and Stoye (2019). This is a generic function, requiring the specification of a test function (test.fun), as well as the specification of the parameter space (hyperparams).

Usage

EAM(
  dir,
  test.fun,
  hyperparams,
  evaluations = NULL,
  time.run.duration = FALSE,
  verbose = 0,
  picturose = FALSE
)

Value

List containing the evaluations of the test statistic and critical values, convergence information, and run times.

Arguments

dir: The direction of the test. dir = 1 corresponds to finding the upper bound of the identified set, dir = -1 corresponds to finding the lower bound.
test.fun: The test function to be inverted in order to obtain the identified set. It should take a scalar parameter as argument (i.e. the specified value of a component of the full parameter vector) and return a list with named elements list(theta, t.stat, crit.val), where theta is the scalar value that was tested, t.stat is the value of the test statistic and crit.val is the critical value to be used in determining whether to reject or not reject.
hyperparams: A list of hyperparameters needed in order for this method to run (see set.EAM.hyperparameters.R).
evaluations: Matrix of already evaluated points, of which at least one is feasible. When evaluations = NULL (default), the initial feasible point search will be executed first.
time.run.duration: Boolean value indicating whether to time each step in the EAM algorithm. Requires chronometer.R. Default is time.run.duration = FALSE.
verbose: Boolean value indicating whether or not to print run time updates to the console. Default is verbose = FALSE.
picturose: Boolean value indicating whether or not to visualize the identified set search. Default is picturose = FALSE.

References

Kaido et al. (2019). Confidence intervals for projections of partially identified parameters. Econometrica. 87(4):1397-1432.