Compute loss function in a flexible way — loss

Compute loss associated with a vector of structural parameters between some data and a transformation of structural parameters

loss_function(
  theta,
  prediction_function = {     function(theta) theta },
  weights = 1L,
  verbose = FALSE,
  return_moment = FALSE,
  moments_weights = NULL,
  moments_weighting_formula = "w ~ moments_weights",
  ...
)

Arguments

theta	Vector of structural parameters. Assuming a named vector.
prediction_function	Function that transforms \(\theta\) into vector of moments. In Newey and MacFadden, denoted \(g(\theta)\)
weights	Weight matrix \(W\) that should be used
verbose	Logical value indicating whether we want to print informative messages to monitor progress
return_moment	Logical value indicating whether we want to return moment matrix (observed, simulation, difference)
moments_weights	User-defined weights that should be applied to reweight moments importance. This is a user choice independent of using an optimal weight matrix
moments_weighting_formula	When relevent, how moment weighting should enter in the objective function
...	Additional arguments that should be used to control `prediction_function` behavior. This function should return a `data.table` object with a variable denoted `epsilon` giving the distance that should be minimized

Value

A distance between empirical and simulated moments

estimation_theta for GMM estimation