Bayesian fused lasso modeling via horseshoe prior

Bayesian fused lasso is one of the sparse Bayesian methods, which shrinks both regression coefficients and their successive differences simultaneously. In this paper, we propose a Bayesian fused lasso modeling via horseshoe prior. By assuming a horseshoe prior on the difference of successive regression coefficients, the proposed method enables us to prevent over-shrinkage of those differences. We also propose a Bayesian nearly hexagonal operator for regression with shrinkage and equality selection with horseshoe prior, which imposes priors on all combinations of differences of regression coefficients. Simulation studies and an application to real data show that the proposed method gives better performance than existing methods.

Automated summary

Bayesian fused lasso is improved by using horseshoe prior on the difference of successive regression coefficients, preventing over-shrinkage. Additionally, a Bayesian nearly hexagonal operator is proposed for regression with shrinkage and equality selection, improving performance compared to existing methods in simulation and real data.