Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: lpoly and weighted regression


From   Umair Khalil <khalil.umair@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: lpoly and weighted regression
Date   Sun, 1 Apr 2012 02:21:31 -0400

Hi,

I had a small question about the lpoly command I used it with the relevant
options:

lpoly Y X if X >0, bw(i) deg(k) kernel(epan2) nograph n(1) at(X0) gen(`s')
se(`serr')  (where X0 = 0, evaluation at 1 point only)

I had to do something else as well so i tried running the local polynomial
regression by myself, I generated the Kernel weights and then just ran an
OLS which amounts to a weighted regression with weights given by the
kernel. However, my results dont match beyond a certain bandwidth for each
degree. For example, with degree 1 it matches with lpoly results till
bandwidth 3, with degree 2 it matches till bandwidth 4, with degree 3 it
matches till 5 and so on. I was wondering why there is a discrepancy in
these results.

Also I wanted a little more detail on the standard error calculation in the
lpoly command. I referred to the stata reference for lpoly but I was
wondering if it could be in some more detail as to how stata is actually
doing the standard error calculation here. Any help would be greatly
appreciated. Thank you.

Umair.

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index