Group variable selection via convex log-exp-sum penalty with application to a breast cancer survivor study

DOI10.1111/BIOM.12230zbMATH Open1419.62166arXiv1306.4397OpenAlexW2768348023WikidataQ33575286 ScholiaQ33575286MaRDI QIDQ3465722FDOQ3465722

Authors: Zhigeng Geng, Sijian Wang, Menggang Yu, Patrick O. Monahan, Victoria Champion, Grace Wahba 

Publication date: 22 January 2016

Published in: Biometrics (Search for Journal in Brave)

Abstract: In many scientific and engineering applications, covariates are naturally grouped. When the group structures are available among covariates, people are usually interested in identifying both important groups and important variables within the selected groups. Among existing successful group variable selection methods, some methods fail to conduct the within group selection. Some methods are able to conduct both group and within group selection, but the corresponding objective functions are non-convex. Such a non-convexity may require extra numerical effort. In this paper, we propose a novel Log-Exp-Sum(LES) penalty for group variable selection. The LES penalty is strictly convex. It can identify important groups as well as select important variables within the group. We develop an efficient group-level coordinate descent algorithm to fit the model. We also derive non-asymptotic error bounds and asymptotic group selection consistency for our method in the high-dimensional setting where the number of covariates can be much larger than the sample size. Numerical results demonstrate the good performance of our method in both variable selection and prediction. We applied the proposed method to an American Cancer Society breast cancer survivor dataset. The findings are clinically meaningful and lead immediately to testable clinical hypotheses.

Full work available at URL: https://arxiv.org/abs/1306.4397

Recommendations

• Group selection in the Cox model with a diverging number of covariates
• Hierarchically penalized Cox regression with grouped variables
• Penalized methods for bi-level variable selection
• Analysis of survival data with group lasso
• Group descent algorithms for nonconvex penalized linear and logistic regression models with grouped predictors

zbMATH Keywords

high-dimensional datapenalized estimationgroup variable selectionsparsity recoverybreast cancer survivorfinite sample bound

Mathematics Subject Classification ID

Linear regression; mixed models (62J05) Applications of statistics to biology and medical sciences; meta analysis (62P10)

Cites Work

• Nearly unbiased variable selection under minimax concave penalty
• A unified approach to model selection and sparse recovery using regularized least squares
• Estimating the dimension of a model
• Pathwise coordinate optimization
• Coordinate descent algorithms for lasso penalized regression
• Variable Selection via Nonconcave Penalized Likelihood and its Oracle Properties
• Title not available (Why is that?)
• Title not available (Why is that?)
• Convergence of a block coordinate descent method for nondifferentiable minimization
• A comparison of GCV and GML for choosing the smoothing parameter in the generalized spline smoothing problem
• Smoothing noisy data with spline functions: Estimating the correct degree of smoothing by the method of generalized cross-validation
• Simultaneous analysis of Lasso and Dantzig selector
• Regularization and Variable Selection Via the Elastic Net
• Model Selection and Estimation in Regression with Grouped Variables
• Can the strengths of AIC and BIC be shared? A conflict between model indentification and regression estimation
• Tuning parameter selectors for the smoothly clipped absolute deviation method
• A group bridge approach for variable selection
• Likelihood-based selection and sharp parameter estimation
• A note on the Lasso and related procedures in model selection
• On the asymptotic properties of the group lasso estimator for linear models
• A fast ``Monte-Carlo cross-validation procedure for large least squares problems with noisy data
• A Stochastic Estimator of the Trace of the Influence Matrix for Laplacian Smoothing Splines

Cited In (3)

• On Data Integration Problems With Manifolds
• Bayesian group bridge for bi-level variable selection
• Grouped penalization estimation of the osteoporosis data in the traditional Chinese medicine

This page was built for publication: Group variable selection via convex log-exp-sum penalty with application to a breast cancer survivor study

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3465722)