Kernel-weighted conditional mean and variance

Useful for the conditional outcome stage of TMLE / AIPW.

Usage

kernel_cond_moments(x, y, x_eval, bandwidth, return_variance = TRUE)

Arguments

x

Numeric covariate vector.

y

Numeric outcome vector.

x_eval

Evaluation grid.

bandwidth

Positive bandwidth.

return_variance

Logical; if FALSE, only the mean is returned.

Value

Either a numeric vector (mean only) or a list with mean and variance.

Examples

set.seed(1)
x <- sort(runif(40, -2, 2))
y <- sin(x) + 0.2 * rnorm(40)
x_eval <- seq(-1.5, 1.5, length.out = 20)
r <- kernel_cond_moments(x, y, x_eval, bandwidth = 0.5,
                         return_variance = TRUE)
r$mean
#>  [1] -0.91406954 -0.88024411 -0.83068530 -0.76548589 -0.68605210 -0.59413791
#>  [7] -0.49015136 -0.37206368 -0.23650553 -0.08252016  0.08401605  0.24988925
#> [13]  0.40055040  0.52769384  0.63127532  0.71593588  0.78660598  0.84619571
#> [19]  0.89546439  0.93412154