{"Package":"mcmcr","Title":"Manipulate MCMC Samples","Version":"0.0.2","Authors@R":"\nperson(\"Joe\", \"Thorley\", email = \"joe@poissonconsulting.ca\", role = c(\"aut\", \"cre\"), comment = c(ORCID = \"0000-0002-7683-4592\"))","Description":"Functions and classes to store, manipulate and summarise\nMonte Carlo Markov Chain (MCMC) samples.","URL":"https://github.com/poissonconsulting/mcmcr","BugReports":"https://github.com/poissonconsulting/mcmcr/issues","License":"MIT + file LICENSE","Depends":"R (>= 2.10)","Imports":"abind, checkr, coda, graphics, stats, tibble, utils","Suggests":"covr, testthat","Encoding":"UTF-8","LazyData":"true","RoxygenNote":"6.1.0","NeedsCompilation":"no","Packaged":"2018-10-30 18:21:14 UTC; joe","Author":"Joe Thorley [aut, cre] ()","Maintainer":"Joe Thorley ","Repository":"CRAN","Date/Publication":"2018-10-30 18:40:02 UTC","repoType":"cran","tarballUrl":"ftp://cran.r-project.org/pub/R/src/contrib/mcmcr_0.0.2.tar.gz","jsonAuthors":[{"name":"Joe Thorley","email":"joe@poissonconsulting.ca","maintainer":true}],"readme":"\n\n\n# mcmcr\n\n[![lifecycle](https://img.shields.io/badge/lifecycle-maturing-blue.svg)](https://www.tidyverse.org/lifecycle/#maturing)\n[![Travis-CI Build\nStatus](https://travis-ci.org/poissonconsulting/mcmcr.svg?branch=master)](https://travis-ci.org/poissonconsulting/mcmcr)\n[![Coverage\nStatus](https://img.shields.io/codecov/c/github/poissonconsulting/mcmcr/master.svg)](https://codecov.io/github/poissonconsulting/mcmcr?branch=master)\n[![License:\nMIT](https://img.shields.io/badge/License-MIT-green.svg)](https://opensource.org/licenses/MIT)\n[![CRAN\\_Status\\_Badge](http://www.r-pkg.org/badges/version/mcmcr)](https://cran.r-project.org/package=mcmcr)\n[![CRAN\nDownloads](http://cranlogs.r-pkg.org/badges/grand-total/mcmcr)](https://CRAN.R-project.org/package=mcmcr)\n[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.1218026.svg)](https://doi.org/10.5281/zenodo.1218026)\n\n## Introduction\n\n`mcmcr` is an R package to manipulate Monte Carlo Markov Chain (MCMC)\nsamples.\n\nFor the purposes of this discussion, an MCMC *sample* represents the\nvalue of a *term* from a single *iteration* of a single *chain*. While a\nsimple *parameter* such as an intercept corresponds to a single term,\nmore complex parameters such as an interaction between two factors\nconsists of multiple terms with their own inherent dimensionality - in\nthis case a matrix. A set of MCMC samples can be stored in different\nways.\n\n### Existing Classes\n\nThe three most common S3 classes store MCMC samples as follows:\n\n - `coda::mcmc` stores the MCMC samples from a single chain as a matrix\n where each each row represents an iteration and each column\n represents a variable\n - `coda::mcmc.list` stores multiple `mcmc` objects (with identical\n dimensions) as a list where each object represents a parallel chain\n - `rjags::mcarray` stores the samples from a single parameter where\n the initial dimensions are the parameter dimensions, the second to\n last dimension is iterations and the last dimension is chains.\n\nIn the first two cases the terms/parameters are represented by a single\ndimension which means that the dimensionality inherent in the parameters\nis stored in the labelling of the variables, ie, `\"bIntercept\",\n\"bInteraction[1,2]\", \"bInteraction[2,1]\", ...`. The structure of the\n`mcmc` and `mcmc.list` objects emphasizes the time-series nature of MCMC\nsamples and is optimized for thining. In contrast `mcarray` objects\npreserve the dimensionality of the parameters.\n\n### New Classes\n\nThe `mcmcr` package defines three related S3 classes which also preserve\nthe dimensionality of the parameters:\n\n - `mcmcr::mcmcarray` is very similar to `rjags::mcarray` except that\n the first dimension is the chains, the second dimension is\n iterations and the subsequent dimensions represent the\n dimensionality of the parameter (it is called `mcmcarray` to\n emphasize that the MCMC dimensions ie the chains and iterations come\n first);\n - `mcmcr::mcmcr` stores multiple uniquely named `mcmcarray` objects\n with the same number of chains and iterations.\n - `mcmcr::mcmcrs` stores multiple `mcmcr` objects with the same\n parameters, chains and iterations.\n\n## Why mcmcr?\n\n`mcmcarray` objects were developed to facilitate manipulation of the\nMCMC samples (although they are just one `aperm` away from `mcarray`\nobjects they are more intuitive to program with - at least for this\nprogrammer\\!). `mcmcr` objects were developed to allow a set of\ndimensionality preserving parameters from a single analysis to be\nmanipulated as a whole. `mcmcrs` objects were developed to allow the\nresults of multiple analyses using the same model to be manipulated\ntogether.\n\nIn addition the `mcmcr` package defines the `term` vector to store and\nmanipulate the term labels, ie, `\"bIntercept\", \"bInteraction[1,2]\",\n\"bInteraction[2,1]\"`, when the MCMC samples are summarised in tabular\nform.\n\nThe `mcmcr` package also introduces a variety of (often) generic\nfunctions to manipulate and query `mcmcarray`, `mcmcr` and `mcmcrs`\nobjects. In particular it provides functions to\n\n - coerce from and to `mcarray`, `mcmc` and `mcmc.list` objects;\n - extract an objects `coef` table (as a tibble);\n - query an object’s `nchains`, `niters`, `npars`, `nterms`, `nsims`\n and `nsams` as well as it’s parameter dimensions (`pdims`) and term\n dimensions (`tdims`);\n - `subset` objects by chains, iterations and/or parameters;\n - `bind_xx` a pair of objects by their `xx_chains`, `_iterations`,\n `xx_parameters` or (parameter) `xx_dimensions`;\n - combine the samples of two (or more) MCMC objects using\n `combine_samples` (or `combine_samples_n`) or combine the samples of\n a single MCMC object by reducing its dimensions using\n `combine_dimensions`;\n - `collapse_chains` or `split_chains` an object’s chains;\n - `mcmc_map` over an objects values;\n - assess if an object has `converged` using `rhat` and `esr`\n (effectively sampling rate);\n - and of course `thin`, `rhat`, `ess` (effective sample size),\n `print`, `plot` etc said objects.\n\nThe code is opinionated which has the advantage of providing a small set\nof stream-lined functions. For example the only ‘convergence’ metric is\nthe uncorrected, untransformed, univariate split R-hat (potential scale\nreduction factor). If you can convince me that additional features are\nimportant I will add them or accept a pull request (see below).\nAlternatively you might want to use the `mcmcr` package to manipulate\nyour samples before coercing them to an `mcmc.list` to take advantage of\nall the summary functions in packages such as `coda`.\n\n## Demonstration\n\n``` r\nlibrary(mcmcr)\n\nmcmcr_example\n#> $alpha\n#> [1] 3.718025 4.718025\n#> \n#> nchains: 2 \n#> niters: 400 \n#> \n#> $beta\n#> [,1] [,2]\n#> [1,] 0.9716535 1.971654\n#> [2,] 1.9716535 2.971654\n#> \n#> nchains: 2 \n#> niters: 400 \n#> \n#> $sigma\n#> [1] 0.7911975\n#> \n#> nchains: 2 \n#> niters: 400\n\ncoef(mcmcr_example)\n#> # A tibble: 7 x 7\n#> term estimate sd zscore lower upper pvalue\n#> \n#> 1 alpha[1] 3.72 0.901 4.15 2.21 5.23 0.00120\n#> 2 alpha[2] 4.72 0.901 5.26 3.21 6.23 0.00120\n#> 3 beta[1,1] 0.972 0.375 2.57 0.251 1.71 0.0225 \n#> 4 beta[2,1] 1.97 0.375 5.24 1.25 2.71 0.005 \n#> 5 beta[1,2] 1.97 0.375 5.24 1.25 2.71 0.005 \n#> 6 beta[2,2] 2.97 0.375 7.91 2.25 3.71 0.00120\n#> 7 sigma 0.791 0.741 1.31 0.425 2.56 0.00120\nrhat(mcmcr_example, by = \"term\")\n#> $alpha\n#> [1] 2.002 2.002\n#> \n#> $beta\n#> [,1] [,2]\n#> [1,] 1.147 1.147\n#> [2,] 1.147 1.147\n#> \n#> $sigma\n#> [1] 0.998\nplot(mcmcr_example[[\"alpha\"]])\n```\n\n

\n\n## Installation\n\nTo install the latest official release from\n[CRAN](https://CRAN.R-project.org/package=mcmcr)\n\n install.packages(\"mcmcr\")\n\nTo install the latest development version from\n[GitHub](https://github.com/poissonconsulting/mcmcr)\n\n # install.packages(\"devtools\")\n devtools::install_github(\"poissonconsulting/err\")\n devtools::install_github(\"poissonconsulting/checkr\")\n devtools::install_github(\"poissonconsulting/mcmcr\")\n\nAlternatively install the latest development version from the Poisson\ndrat [repository](https://github.com/poissonconsulting/drat)\n\n # install.packages(\"drat\")\n drat::addRepo(\"poissonconsulting\")\n install.packages(\"mcmcr\")\n\n## Citation\n\n``` \n\nTo cite package 'mcmcr' in publications use:\n\n Joe Thorley (2018). mcmcr: Manipulate MCMC Samples. R package\n version 0.0.1.9002. https://github.com/poissonconsulting/mcmcr\n\nA BibTeX entry for LaTeX users is\n\n @Manual{,\n title = {mcmcr: Manipulate MCMC Samples},\n author = {Joe Thorley},\n year = {2018},\n note = {R package version 0.0.1.9002},\n url = {https://github.com/poissonconsulting/mcmcr},\n }\n```\n\n## Contribution\n\nPlease report any\n[issues](https://github.com/poissonconsulting/mcmcr/issues).\n\n[Pull requests](https://github.com/poissonconsulting/mcmcr/pulls) are\nalways welcome.\n\nPlease note that this project is released with a [Contributor Code of\nConduct](CONDUCT.md). By participating in this project you agree to\nabide by its terms.\n\n## Inspiration\n\n[coda](https://github.com/cran/coda) and\n[rjags](https://github.com/cran/rjags)","jobInfo":{"package":"mcmcr","version":"0.0.2","parsingStatus":"success","parserVersion":1,"parsedAt":"2018-11-24T07:29:51+0000"}}

Description

Arguments

Examples