get_est {RcausalEGM} | R Documentation |
Make predictions with causalEGM model.
Description
When x is NULL, the conditional average treatment effect (CATE), namely tau(v), is estimated using a trained causalEGM model. When x is provided, estimating the potential outcome y given treatment x and covariates v using a trained causalEGM model.
Usage
get_est(object, v, x = NULL)
Arguments
object |
An object of class |
v |
is the covariates, two-dimensional array with size n by p. |
x |
is the optional treatment variable, one-dimensional array with size n. Defaults to NULL. |
Value
Vector of predictions.
Examples
#Generate a simple simulation data.
n <- 1000
p <- 10
v <- matrix(rnorm(n * p), n, p)
x <- rbinom(n, 1, 0.4 + 0.2 * (v[, 1] > 0))
y <- pmax(v[, 1], 0) * x + v[, 2] + pmin(v[, 3], 0) + rnorm(n)
model <- causalegm(x=x, y=y, v=v, n_iter=3000)
n_test <- 100
v_test <- matrix(rnorm(n_test * p), n_test, p)
x_test <- rbinom(n_test, 1, 0.4 + 0.2 * (v_test[, 1] > 0))
pred_cate <- get_est(model, v = v_test) # CATE estimate
pred_y <- get_est(model, v = v_test, x = x_test) # y given treatment x plus covariates v
[Package RcausalEGM version 0.3.3 Index]