On stepwise pattern recovery of the fused Lasso

DOI10.1016/J.CSDA.2015.08.013zbMATH Open1468.62161arXiv1211.5194OpenAlexW2200384000MaRDI QIDQ1660156FDOQ1660156

Jinzhu Jia, Junyang Qian

Publication date: 15 August 2018

Published in: Computational Statistics and Data Analysis (Search for Journal in Brave)

Abstract: We study the property of the Fused Lasso Signal Approximator (FLSA) for estimating a blocky signal sequence with additive noise. We transform the FLSA to an ordinary Lasso problem. By studying the property of the design matrix in the transformed Lasso problem, we find that the irrepresentable condition might not hold, in which case we show that the FLSA might not be able to recover the signal pattern. We then apply the newly developed preconditioning method -- Puffer Transformation [Jia and Rohe, 2012] on the transformed Lasso problem. We call the new method the preconditioned fused Lasso and we give non-asymptotic results for this method. Results show that when the signal jump strength (signal difference between two neighboring groups) is big and the noise level is small, our preconditioned fused Lasso estimator gives the correct pattern with high probability. Theoretical results give insight on what controls the signal pattern recovery ability -- it is the noise level {instead of} the length of the sequence. Simulations confirm our theorems and show significant improvement of the preconditioned fused Lasso estimator over the vanilla FLSA.

Full work available at URL: https://arxiv.org/abs/1211.5194

Recommendations

• Properties and refinements of the fused Lasso
• Preconditioning the Lasso for sign consistency
• Sparsity and Smoothness Via the Fused Lasso
• Split Bregman method for large scale fused Lasso
• Modified path algorithm of fused Lasso signal approximator for consistent recovery of change points

zbMATH Keywords

preconditioningfused Lassonon-asymptoticpattern recovery

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Ridge regression; shrinkage estimators (Lasso) (62J07)

Cites Work

• Pathwise coordinate optimization
• Variable Selection via Nonconcave Penalized Likelihood and its Oracle Properties
• Title not available (Why is that?)
• High-dimensional graphs and variable selection with the Lasso
• Title not available (Why is that?)
• Sparsity and Smoothness Via the Fused Lasso
• The solution path of the generalized lasso
• Properties and refinements of the fused Lasso
• Persistene in high-dimensional linear predictor-selection and the virtue of overparametrization
• Asymptotics for Lasso-type estimators.
• A Singular Value Thresholding Algorithm for Matrix Completion
• Multiple Change-Point Estimation With a Total Variation Penalty
• The sparsity and bias of the LASSO selection in high-dimensional linear regression
• Stable recovery of sparse overcomplete representations in the presence of noise
• Just relax: convex programming methods for identifying sparse signals in noise
• Spatial smoothing and hot spot detection for CGH data using the fused lasso
• Preconditioning the Lasso for sign consistency
• Title not available (Why is that?)
• Information-Theoretic Limits on Sparsity Recovery in the High-Dimensional and Noisy Setting
• Title not available (Why is that?)

Cited In (15)

• Orthogonal one step greedy procedure for heteroscedastic linear models
• Multiple change-point detection: a selective overview
• A modified information criterion for tuning parameter selection in 1d fused LASSO for inference on multiple change points
• Modified path algorithm of fused Lasso signal approximator for consistent recovery of change points
• Tuning parameter selection in fused lasso signal approximator with false discovery rate control
• Preconditioning the Lasso for sign consistency
• The DFS Fused Lasso: Linear-Time Denoising over General Graphs
• Exact Spike Train Inference Via $\ell_0$ Optimization
• Oracle Efficient Estimation of Structural Breaks in Cointegrating Regressions
• Prediction bounds for higher order total variation regularized least squares
• Path algorithms for fused lasso signal approximator with application to COVID‐19 spread in Korea
• Optimal covariance change point localization in high dimensions
• Empirical priors and posterior concentration in a piecewise polynomial sequence model
• Properties and refinements of the fused Lasso
• On the total variation regularized estimator over a class of tree graphs

This page was built for publication: On stepwise pattern recovery of the fused Lasso

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1660156)