A Cross Validation Framework for Signal Denoising with Applications to Trend Filtering, Dyadic CART and Beyond

TL;DR

This paper introduces a general cross validation framework for signal denoising that applies to methods like Trend Filtering and Dyadic CART, achieving near-optimal convergence rates and extending to other estimators.

Contribution

It provides the first theoretical analysis of cross validated versions of Trend Filtering and Dyadic CART, demonstrating their effectiveness and broad applicability.

Findings

Cross validated methods attain near-optimal convergence rates.

Framework is applicable to high-dimensional linear regression and matrix estimation.

Extends theoretical understanding of cross validation in nonparametric regression.

Abstract

This paper formulates a general cross validation framework for signal denoising. The general framework is then applied to nonparametric regression methods such as Trend Filtering and Dyadic CART. The resulting cross validated versions are then shown to attain nearly the same rates of convergence as are known for the optimally tuned analogues. There did not exist any previous theoretical analyses of cross validated versions of Trend Filtering or Dyadic CART. To illustrate the generality of the framework we also propose and study cross validated versions of two fundamental estimators; lasso for high dimensional linear regression and singular value thresholding for matrix estimation. Our general framework is inspired by the ideas in Chatterjee and Jafarov (2015) and is potentially applicable to a wide range of estimation methods which use tuning parameters.