predict.mvoprobit {switchSelection}

Predict method for mvoprobit function

Description

Predicted values based on object of class 'mvoprobit'.

Usage

## S3 method for class 'mvoprobit'
predict(
  object,
  ...,
  newdata = NULL,
  given_ind = numeric(),
  group = NULL,
  group2 = NULL,
  type = ifelse(is.null(group2), "prob", "val"),
  me = NULL,
  eps = NULL,
  control = list(),
  se = FALSE
)

Arguments

object

object of class "mvoprobit"

...

further arguments (currently ignored)

newdata

an optional data frame in which to look for variables with which to predict. If omitted, the original data frame used. This data frame should contain values of dependent variables even if they are not actually needed for prediction (simply assign them with 0 values).

given_ind

numeric vector of indexes of conditioned components.

group

numeric vector which i-th element represents a value of the i-th dependent variable. If this value equals -1 then this component will be ignored (useful for estimation of marginal probabilities).

group2

numeric vector which i-th element represents a value of the i-th dependent variable of the continuous equation. If this value equals -1 then this component will be ignored.

type

string representing a type of prediction. See 'Details' for more information.

me

string representing the name of the variable for which marginal effect should be estimated. See 'Details' for more information.

eps

numeric vector of length 1 or 2 used for calculation of marginal effects. See 'Details'.

control

list of additional arguments. Currently is not intended for the users.

se

logical; if TRUE then the function also returns standard errors and p-values of the two-sided significance test associated with the function output. Works only if predict.mvoprobit returns numeric vector or a single column matrix. See delta_method for more information.

Details

See 'Examples' section of mvoprobit for examples of this function application.

If type = "prob" then function returns a joint probability that dependent variables will have values assigned in group. To calculate marginal probabilities set unnecessary group values to -1. To estimate conditional probabilities provide indexes of conditioned variables through given_ind. For example if z_{1}, z_{2} and z_{3} are dependent variables then to calculate P(z_{1}=2 | z_{3} = 0) set given_ind = 3 and groups = c(2, -1, 0). Note that conditioning on covariates (regressors) is omitted for notations brevity and this conditioning depends on the values in newdata.

If type = "li" then function returns a matrix which columns are linear indexes of corresponding equations.

If type = "sd" then function returns a matrix which columns are standard deviations of random errors for corresponding equations.

If type = "li" or type = "sd" and some groups are equal to -1 then corresponding components will be omitted from the output matrix.

If type = "val" then function returns predictions of conditional (on group) expectation of dependent variable in continuous equations with regimes determined by group2 argument. To predict unconditional expectations just set group = NULL.

If type = "lambda" then function returns conditional (on group) expectations of random error of continuous equation in regime group2.

If type = "val" or type = "lambda" then output is a matrix which i-th column corresponds to prediction associated with i-th continuous equation.

If me is provided then the function returns marginal effect of variable me respect to the statistic determined by type argument. For example if me = "x1" and type = "prob" then function returns marginal effect of x1 on the corresponding probability i.e. one that would be estimated if me is NULL.

If length(eps) = 1 then eps is an increment in numeric differentiation procedure. If eps is NULL then this increment will be selected automatically taking into account scaling of variables. If length(eps) = 2 then marginal effects will be estimated as the difference of predicted value when variable me equals eps[2] and eps[1] correspondingly.

For example suppose that type = "prob", me = "x1", given_ind = 3 and groups = c(2, -1, 0). Then if eps is a NULL or a small number (something like eps = 0.0001) the following marginal effect will be estimated (via first difference numeric differentiation):

\frac{\partial P(z_{1}=2 | z_{3} = 0)}{\partial x_{1}}.

If eps = c(1, 3) then the function estimates the following difference (useful for estimation of marginal effects of ordered covariates):

P(z_{1}=2 | z_{3} = 0, x_{1} = 3) - P(z_{1}=2 | z_{3} = 0, x_{1} = 1).

Value

This function returns predictions for each row of newdata or for each observation in the model if newdata is NULL. Structure of the output depends on the type argument (see 'Details' section).