|
Activity Number:
|
187
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308436 |
|
Title:
|
Adaptive Regressograms
|
|
Author(s):
|
Anthony Gamst*+
|
|
Companies:
|
University of California, San Diego
|
|
Address:
|
Biostatistics, La Jolla, CA, 92093-0717,
|
|
Keywords:
|
minimax ; rate adaptive ; regression ; penalized least squares ; quasi-likelihood ; model selection
|
|
Abstract:
|
We study the problem of selecting the number and location of break-points for piece-wise constant regression function estimates. When the underlying regression function has smoothness parameter s, it is known that the optimal number of bins increases at the rate 1/(2s+1). Simple model selection penalties lead to regressograms with nearly optimal risk. These estimates also adapt to the unknown smoothness s, leading to optimal rates of convergence regardless of the underlying smoothness. The arguments involved can be extended to other spline bases and to quasi-likelihood-type regression problems.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
