Hi,
I have read Locally Weighted Regression (LOESS) from the
book "Introduction to Linear Regression Analysis" by Montogomery,
Peck and Vining and then I have tried it through SAS. I face
following problems.
1. MODEL dep. Vars = indep. Vars / DIRECT;
In SAS-help file it is written "DIRECT option specifies that local
least squares fits are to be done…". I have found in the book "the
loess procedure uses the points in the neighbourhood to generate a
weighted least square estimate of the specific response" (page no
239, third edition).
Does here the "local least squares" in SAS-help is equivalent
to "Weighted Least Squares" in the book?
2. MODEL dep. Vars = indep. Vars / INTERP=linear/cubic;
It is written is SAS-help "The INTERP= option specifies the degree
of the interpolating polynomials used for blending local polynomial
fits at the kd tree vertices."
I do not get what it tells.
3. How can I specify TRI-CUBE weighting function in LOESS procedure?
4. What does it mean by KD TREE?
If anybody knows the answer then I sincerely request to answer my
questions.
With regards,
Madan
--------------
Madan Gopal Kundu
Room no 18, Sukhatme Hostel
Indian Agricultural Statistics Research Institute
Library Avenue, Pusa
New Delhi 12
India
Web: http://www.freewebs.com/madanstata
mobile: 91-9868788406
e-mail: [log in to unmask]
---------------------------------
Enjoy this Diwali with Y! India Click here
|