Online Program Home
My Program

Abstract Details

Activity Number: 314
Type: Contributed
Date/Time: Tuesday, August 2, 2016 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistical Learning and Data Science
Abstract #320170 View Presentation
Title: Nonparametric Regression with Adaptive Smoothness via a Convex Hierarchical Penalty
Author(s): Asad Haris* and Ali Shojaie and Noah Simon
Companies: University of Washington and University of Washington and University of Washington
Keywords: Nonparametric regression ; Convex optimization ; Big data ; Penalized regression ; Additive models ; Convergence rates

We consider the task of fitting a nonparametric regression model using a growing set of basis functions. Like projection estimators, our proposal fits a truncated basis expansion. The novelty of our method lies in the data-driven mechanism for selection of the truncation level which is achieved by nested group penalties. We establish convergence rates comparable to smoothing splines via entropy methods and, a closed form solution for the resulting optimization problem. We explore its performance in a simulation study and on real data.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2016 program

Copyright © American Statistical Association