Usage: |
mh = lregxestp(x {,h {,v} })
|
Input: |
| x | n x (k+1), the data. The independent variables are
contained in the first k columns and the dependent
variable in the last one.
|
| h | optional scalar or k x 1 vector of bandwidth. If not
given, 20% of the volume of x[,1:k] is used as default.
|
| v | optional m x k matrix, values of the independent variable
in which to compute the regression.
If not given and k < 4, a grid of length 100 (k = 1),
length 30 (k = 2) or length 8 (k = 3) is used.
If k >= 4 then v is set to (sorted) x.
|
Output: |
| mh | n x (k+1) or m x (k+1) matrix; the first k columns
contain the sorted x[,1:k], the sorted grid or
the sorted v and the last column contains the
regression estimate given at the values of the
first k columns. |