{"Package":"Mhorseshoe","Title":"Approximate Algorithm for Horseshoe Prior","Version":"0.1.5","Authors@R":"c(\nperson(\"Kang\", \"Mingi\", , \"leehuimin115@g.skku.edu\", role = c(\"aut\",\"cre\")),\nperson(\"Lee\", \"Kyoungjae\", , \"leekjstat@gmail.com\", role = \"aut\"))","Description":"Provides exact and approximate algorithms for the horseshoe prior\nin linear regression models, which were proposed by Johndrow et al. (2020)\n.","Encoding":"UTF-8","Imports":"stats, Rcpp (>= 1.0.11)","LinkingTo":"Rcpp","RoxygenNote":"7.3.3","License":"MIT + file LICENSE","Suggests":"knitr, rmarkdown, ggplot2, testthat (>= 3.0.0)","VignetteBuilder":"knitr","Config/testthat/edition":"3","NeedsCompilation":"yes","Packaged":"2025-10-20 09:43:10 UTC; rkdal","Author":"Kang Mingi [aut, cre],\nLee Kyoungjae [aut]","Maintainer":"Kang Mingi ","Repository":"CRAN","Date/Publication":"2025-10-20 10:00:02 UTC","repoType":"cran","tarballUrl":"ftp://cran.r-project.org/pub/R/src/contrib/Mhorseshoe_0.1.5.tar.gz","jsonAuthors":[{"name":"Kang Mingi","email":"leehuimin115@g.skku.edu","maintainer":true},{"name":"Lee Kyoungjae","email":"leekjstat@gmail.com","maintainer":false}],"readme":"# Mhorseshoe\n\n## Overview\n\nMhorseshoe is a package for a high-dimensional Bayesian linear modeling \nalgorithm using a horseshoe prior. A feature of this package is that it \nimplements approximate MCMC algorithm from Johndrow et al. (2020) and provides \na horseshoe estimator that can effectively reduce computational costs for \nhigh-dimensional sparse data. This package provides two different algorithm \nfunctions :\n\n-`exact_horseshoe()` Run the horseshoe estimator.\n\n-`approx_horseshoe()` Run the horseshoe estimator with the approximate algorithm applied.\n\n## Installation\n\n```r\ninstall.package(\"Mhorseshoe\")\n```\n\n## Usage\n\nThe following linear model assumptions are made.\n\n$$L(y\\ |\\ x, \\beta, \\sigma^2) = (\\frac{1}{\\sqrt{2\\pi}\\sigma})^{-N/2}exp\n\\{ -\\frac{1}{2\\sigma^2}(y-X\\beta)^T(y-X\\beta)\\},\\\\ X \\in \n\\mathbb{R}^{N \\times p},\\ y \\in \\mathbb{R}^{N},\\ \\beta \\in \\mathbb{R}^{p}$$\n\n- $X \\in \\mathbb{R}^{N \\times p}$ : Matrix of covariates.\n- $y \\in \\mathbb{R}^{N}$ : Response variable.\n\n```r\n# Run functions from the Mhorseshoe package\nex_result <- exact_horseshoe(y, X, burn = 5000, iter = 10000)\nap_result <- approx_horseshoe(y, X, burn = 5000, iter = 10000)\n\n# posterior mean of beta\nex_betahat <- ex_result$BetaHat\nap_betahat <- ap_result$BetaHat\n\n# 95% posterior credible intervals\nex_LeftCI <- ex_result$LeftCI\nex_RightCI <- ex_result$RightCI\nap_LeftCI <- ap_result$LeftCI\nap_RightCI <- ap_result$RightCI\n```\n\n## References\n\nJohndrow, J., Orenstein, P., & Bhattacharya, A. (2020). Scalable Approximate \nMCMC Algorithms for the Horseshoe Prior. In Journal of Machine Learning \nResearch (Vol. 21).\n\nIf you would like to discuss this package, please email leehuimin115@g.skku.edu","jobInfo":{"package":"Mhorseshoe","version":"0.1.5","parsingStatus":"success","parserVersion":2,"parsedAt":"2025-10-20T12:30:22+0000"}}

Description

Usage

Arguments