{"Package":"ScaleSpikeSlab","Type":"Package","Title":"Scalable Spike-and-Slab","Version":"1.0","Date":"2022-05-13","Authors@R":"\nc(person(given = \"Niloy\",\nfamily = \"Biswas\",\nrole = c(\"aut\", \"cre\"),\nemail = \"niloy_biswas@g.harvard.edu\",\ncomment = c(ORCID = \"0000-0001-9081-5702\")),\nperson(given = \"Lester\",\nfamily = \"Mackey\",\nrole = \"aut\"),\nperson(given = \"Xiao-Li\",\nfamily = \"Meng\",\nrole = \"aut\"))","Description":"A scalable Gibbs sampling implementation for high dimensional Bayesian regression with the continuous spike-and-slab prior. Niloy Biswas, Lester Mackey and Xiao-Li Meng, \"Scalable Spike-and-Slab\" (2022) .","License":"GPL (>= 2)","Imports":"Rcpp, stats, TruncatedNormal","LinkingTo":"Rcpp, RcppEigen","RoxygenNote":"7.1.2","NeedsCompilation":"yes","Packaged":"2022-05-17 20:40:23 UTC; niloybiswas","Author":"Niloy Biswas [aut, cre] (),\nLester Mackey [aut],\nXiao-Li Meng [aut]","Maintainer":"Niloy Biswas ","Depends":"R (>= 3.5.0)","Repository":"CRAN","Date/Publication":"2022-05-18 17:00:07 UTC","repoType":"cran","tarballUrl":"ftp://cran.r-project.org/pub/R/src/contrib/ScaleSpikeSlab_1.0.tar.gz","jsonAuthors":[{"name":"Niloy Biswas","email":"niloy_biswas@g.harvard.edu","maintainer":true},{"name":"Lester Mackey","email":null,"maintainer":false},{"name":"Xiao-Li Meng","email":null,"maintainer":false}],"readme":"---\ntitle: \"ScaleSpikeSlab\"\nauthor: \"Niloy Biswas\"\ndate: \"31/01/2022\"\noutput: \n html_document: \n keep_md: yes\n---\n\n\n\n# ScaleSpikeSlab (S^3)\n\nThis package contains algorithms for *Scalable Spike-and-Slab* (S^3),\na scalable Gibbs sampling implementation for high-dimensional Bayesian \nregression with the continuous spike-and-slab prior.\n\nIt is based on the article \"Scalable Spike-and-Slab\", \nby Niloy Biswas, Lester Mackey and Xiao-Li Meng. The folder `inst` contains \nscripts to reproduce the results of the article.\n\n## Installation\n\nThe package can be installed from R via:\n\n\n```r\n# install.packages(\"devtools\")\ndevtools::install_github(\"niloyb/ScaleSpikeSlab/R_package\")\n\n# Install dependencies Rcpp, RcppEigen\ninstall.packages(c(\"Rcpp\", \"RcppEigen\"))\n# Install additional packages to help with parallel computation and plotting\ninstall.packages(c(\"doParallel\", \"doRNG\", \"foreach\", \"dplyr\", \"tidyr\", \n \"ggplot2\", \"latex2exp\", \"reshape2\", \"ggpubr\"))\n```\n\n## A tutorial with Riboflavin GWAS data\n\n#### Import data and select hyperparameters. \n\n```r\nset.seed(1)\nlibrary(ScaleSpikeSlab)\n\n# Riboflavin linear regression dataset of Buhlmann et al. (2014) \ndata(riboflavin)\nX <- riboflavin$x\nXt <- t(X)\ny <- riboflavin$y\n\n# Choose hyperparamters\nparams <- spike_slab_params(n=nrow(X),p=ncol(X))\n```\n#### Run MCMC with S^3\n\n\n```r\nlibrary(doParallel)\nregisterDoParallel(cores = detectCores()-1)\nlibrary(foreach)\n\nno_chains <- 50\nsss_chain_z_output <- \n foreach(i = c(1:no_chains), .combine=rbind)%dopar%{\n sss_chain <- spike_slab_linear(chain_length=5e3,burnin=1e3,X=X,Xt=Xt,y=y,\n tau0=params$tau0,tau1=params$tau1,q=params$q,\n verbose=FALSE,store=FALSE)\n return(as.vector(sss_chain$z_ergodic_avg))\n}\n```\n#### Plot Spike-and-Slab marginal posterior probabilities for variable selection \n\n\n```r\nlibrary(dplyr)\nlibrary(ggplot2)\nlibrary(latex2exp)\n\nriboflavin_df <- \n data.frame(post_prob_mean=apply(sss_chain_z_output,2,mean),\n post_prob_sd=apply(sss_chain_z_output,2,sd),\n cov_index=c(1:ncol(X)), no_chains=no_chains) %>%\n arrange(desc(post_prob_mean)) %>%\n mutate(xaxis =1:n())\n\n\nggplot(riboflavin_df, aes(x=xaxis, y=post_prob_mean)) + \n geom_point(size=2) + \n geom_errorbar(aes(ymax=(post_prob_mean+3*post_prob_sd/sqrt(no_chains)), \n ymin=(post_prob_mean-3*post_prob_sd/sqrt(no_chains))),\n position=position_dodge(.9)) +\n xlab('Riboflavin Covariates') + \n ylab(TeX('Marginal posterior probabilities')) +\n scale_x_continuous(trans='log10') + theme_classic(base_size = 12)\n```\n\n![](README_files/figure-html/plot_sss-1.png)","jobInfo":{"package":"ScaleSpikeSlab","version":"1.0","parsingStatus":"success","parserVersion":1,"parsedAt":"2022-05-18T19:29:53+0000"}}

Description

Usage

Arguments

Value

Examples