check.lpmodel: Check function: 'lpmodel'
In conroylau/lpinfer: An R Package for Inference in Linear Programs

check.lpmodel

R Documentation

Check function: `lpmodel`

This function checks if the object lpmodel is in the correct format.

check.lpmodel(
  data,
  lpmodel,
  name.var,
  A.tgt.cat,
  A.obs.cat,
  A.shp.cat,
  beta.obs.cat,
  beta.shp.cat,
  R,
  is.estbounds = FALSE
)

`data`	An `data frame` or a `matrix`.
`lpmodel`	An `lpmodel` object.
`name.var`	The name of the `lpmodel` object.
`A.tgt.cat`	The category of the `A.tgt` object.
`A.obs.cat`	The category of the `A.obs` object.
`A.shp.cat`	The category of the `A.shp` object.
`beta.obs.cat`	The category of the `beta.obs` object.
`beta.shp.cat`	The category of the `beta.shp` object.
`R`	The number of bootstrap replications.
`is.estbounds`	A boolean variable that indicates whether the test function being called is `estbounds`. If this function is being called, then it means that when `data` is `NULL`, then each component of the `lpmodel` object is not required to be a `list`. Otherwise, the `lpmodel` object needs to contain the bootstrap estimates if `data` is `NULL`.

In each of the testing procedures, there are six possible categories for the each of the objects:

not_used: This refers to the case where the object is not used in the function.
matrix: This refers to the case where the object has to be a matrix.
function_mat: This refers to the case where the object has to be a function that produces a matrix.
list: This refers to the case where the object is a list.
function_obs_var: This refers to the case where the object is a function that produces a list that contains a matrix and a vector. This is typically the case for beta.obs when the testing procedure requires both the observed value of beta.obs and the estimator of the asymptotic variance.
function_obs_var_bs: This is essentially the same as the function_obs_var category. However, when the variance matrix is not provided, it will be estimated by standard nonparametric bootstrap so there is no problem in case the variance matrix is not provided.