Library: | metrics |
See also: | adeslp dwade trimper wtsder |
Quantlet: | adeind | |
Description: | indirect average derivative estimation using binning |
Usage: | {delta,dvar} = adeind(x,y,d,m) | |
Input: | ||
x | n x p matrix , the observed explanatory variable | |
y | n x 1 matrix , the observed response variable | |
d | p x 1 vector or scalar , the binwidth or the grid | |
m | p x 1 vector or scalar , the bandwidth to be used during estimation of the scores | |
Output: | ||
delta | p x 1 vector , the ADE estimate | |
dvar | p x p matrix , the estimated asymptotic covariance matrix of delta |
library("metrics") randomize(0) n = 100 x = normal(n,3) z = 0.2*x[,1] - 0.7*x[,2] + x[,3] eps = normal(n,1) * sqrt(0.5) y = 2 * z^3 + eps d = 0.2 m = 5 {delta,dvar} = adeind(x,y,d,m) delta dvar
the indirect regression estimator for average derivative and its asymtotic covariance matrix as described by Haerdle and Stoker, JASA (1989) and Turlach, Discussion Paper (1993)