Bayesian L_1/2 regression

08/07/2021
by   Xiongwen Ke, et al.
0

It is well known that bridge regression enjoys superior theoretical properties than traditional LASSO. However, the current latent variable representation of its Bayesian counterpart, based on the exponential power prior, is computationally expensive in higher dimensions. In this paper, we show that the exponential power prior has a closed-form scale mixture of normal decomposition for α=(1/2)^γ, γ∈ℕ^+. We develop a partially collapsed Gibbs sampling scheme, which outperforms existing Markov chain Monte Carlo strategies, we also study theoretical properties under this prior when p>n. In addition, we introduce a non-separable bridge penalty function inspired by the fully Bayesian formulation and a novel, efficient, coordinate-descent algorithm. We prove the algorithm's convergence and show that the local minimizer from our optimization algorithm has an oracle property. Finally, simulation studies were carried out to illustrate the performance of the new algorithms.

READ FULL TEXT
research
04/21/2021

Precision Matrix Estimation under the Horseshoe-like Prior-Penalty Dual

The problem of precision matrix estimation in a multivariate Gaussian mo...
research
12/06/2022

SURE-tuned Bridge Regression

Consider the ℓ_α regularized linear regression, also termed Bridge regre...
research
09/29/2022

Bayesian Quantile Regression for Ordinal Models

The paper introduces a Bayesian estimation method for quantile regressio...
research
11/30/2018

Markov chain Monte Carlo Methods For Lattice Gaussian Sampling:Convergence Analysis and Enhancement

Sampling from lattice Gaussian distribution has emerged as an important ...
research
02/09/2020

Stochastic tree ensembles for regularized nonlinear regression

This paper develops a novel stochastic tree ensemble method for nonlinea...
research
01/31/2023

Classified as unknown: A novel Bayesian neural network

We establish estimations for the parameters of the output distribution f...
research
11/12/2022

A Credibility Index Approach for Effective a Posteriori Ratemaking with Large Insurance Portfolios

Credibility, experience rating and more recently the so-called a posteri...

Please sign up or login with your details

Forgot password? Click here to reset