sigma2.NN {FRD} | R Documentation |
Fast routine to calculate conditional variance based on nearest neighbours based on linear regression.
sigma2.NN(x, d, y, R = 5, type = "regression NN")
x |
Running variable. Note that the entries of this vector have to be sorted. |
d |
Treatment status. |
y |
Outcome variable. |
R |
Number of neares neighbours to use. Default is 5. |
type |
Either "regression NN" calculates variance based on local linear projections of R-nearest neighbours or "mean NN" calculates conditional variance based on mean of R-nearest neighbours. |
This function calculates local conditional variances. The output is a nx4 matrix. The first column equals$\sigma_y^2$, the second and third columns $\sigma_YT$ and the forth column $\sigma_T^2$