{"Package":"Rdimtools","Type":"Package","Title":"Dimension Reduction and Estimation Methods","Version":"1.0.4","Authors@R":"c(person(\"Kisung\", \"You\", role = c(\"aut\", \"cre\"), email = \"kyoustat@gmail.com\",comment=c(ORCID=\"0000-0002-8584-459X\")), person(\"Changhee\",\"Suh\",role=c(\"ctb\"),email=\"cs19bc@my.fsu.edu\"))","Description":"We provide linear and nonlinear dimension reduction techniques.\nIntrinsic dimension estimation methods for exploratory analysis are also provided.\nFor more details on the package, see the paper by You (2020) .","License":"MIT + file LICENSE","Encoding":"UTF-8","LazyData":"true","Depends":"R (>= 3.0.0)","Imports":"CVXR (>= 1.0), Rcpp (>= 0.12.15), RcppDE, Rcsdp, Rdpack,\nRSpectra, graphics, maotai (>= 0.1.6), mclustcomp, stats, utils","LinkingTo":"Rcpp, RcppArmadillo, RcppDist, maotai","RoxygenNote":"7.1.1","RdMacros":"Rdpack","URL":"http://kyoustat.com/Rdimtools/","BugReports":"https://github.com/kyoustat/Rdimtools/issues","Suggests":"knitr, rmarkdown","VignetteBuilder":"knitr","NeedsCompilation":"yes","Packaged":"2020-09-07 21:20:31 UTC; kisung","Author":"Kisung You [aut, cre] (),\nChanghee Suh [ctb]","Maintainer":"Kisung You ","Repository":"CRAN","Date/Publication":"2020-09-07 23:21:37","repoType":"cran","tarballUrl":"ftp://cran.r-project.org/pub/R/src/contrib/Rdimtools_1.0.4.tar.gz","jsonAuthors":[{"name":"Kisung You","email":"kyoustat@gmail.com","maintainer":true},{"name":"Changhee Suh","email":"cs19bc@my.fsu.edu","maintainer":false}],"readme":"\n\n\n# Rdimtools

\n\n\n\n[![CRAN\\_Status\\_Badge](http://www.r-pkg.org/badges/version/Rdimtools?color=green)](https://cran.r-project.org/package=Rdimtools)\n[![Travis-CI Build\nStatus](https://travis-ci.org/kyoustat/Rdimtools.svg?branch=master)](https://travis-ci.org/kyoustat/Rdimtools)\n[![](https://cranlogs.r-pkg.org/badges/Rdimtools)](https://cran.r-project.org/package=Rdimtools)\n\n\n**Rdimtools** is an R package for dimension reduction (DR) - including\nfeature selection and manifold learning - and intrnsic dimension\nestimation (IDE) methods. We aim at building one of the *most\ncomprehensive* toolbox available online, where current version delivers\n139 DR algorithms and 17 IDE methods.\n\nThe philosophy is simple, **the more we have at hands, the better we can\nplay**.\n\n## Elephant\n\nOur logo characterizes the foundational nature of multivariate data\nanalysis; we may be blind people wrangling the data to see an\n[elephant](https://en.wikipedia.org/wiki/Blind_men_and_an_elephant) to\ngrasp an idea of what the data looks like with partial information from\neach algorithm.\n\n## Installation\n\nYou can install a release version from CRAN:\n\n``` r\ninstall.packages(\"Rdimtools\")\n```\n\nor the development version from github:\n\n``` r\n## install.packages(\"devtools\")\ndevtools::install_github(\"kyoustat/Rdimtools\")\n```\n\n## Minimal Example : Dimension Reduction\n\nHere is an example of dimension reduction on the famous `iris` dataset.\nPrincipal Component Analysis (`do.pca`), Laplacian Score (`do.lscore`),\nand Diffusion Maps (`do.dm`) are compared, each from a family of\nalgorithms for linear reduction, feature extraction, and nonlinear\nreduction.\n\n``` r\n# load the library\nlibrary(Rdimtools)\n\n# load the data\nX = as.matrix(iris[,1:4])\nlab = as.factor(iris[,5])\n\n# run 3 algorithms mentioned above\nmypca = do.pca(X, ndim=2)\nmylap = do.lscore(X, ndim=2)\nmydfm = do.dm(X, ndim=2, bandwidth=10)\n\n# visualize\npar(mfrow=c(1,3))\nplot(mypca$Y, pch=19, col=lab, xlab=\"axis 1\", ylab=\"axis 2\", main=\"PCA\")\nplot(mylap$Y, pch=19, col=lab, xlab=\"axis 1\", ylab=\"axis 2\", main=\"Laplacian Score\")\nplot(mydfm$Y, pch=19, col=lab, xlab=\"axis 1\", ylab=\"axis 2\", main=\"Diffusion Maps\")\n```\n\n

\n\n## Minimal Example : Dimension Estimation\n\n![](http://people.cs.uchicago.edu/~dinoj/manifold/swissroll.gif)\n\nSwill Roll is a classic example of 2-dimensional manifold embedded in\n\$\\mathbb{R}^3\$ and one of 11 famous model-based samples from\n`aux.gensamples()` function. Given the ground truth that \$d=2\$, let’s\napply several methods for intrinsic dimension estimation.\n\n``` r\n# generate sample data\nset.seed(100)\nroll = aux.gensamples(dname=\"swiss\")\n\n# we will compare 6 methods (out of 17 methods from version 1.0.0)\nvecd = rep(0,5)\nvecd[1] = est.Ustat(roll)$estdim # convergence rate of U-statistic on manifold\nvecd[2] = est.correlation(roll)$estdim # correlation dimension\nvecd[3] = est.made(roll)$estdim # manifold-adaptive dimension estimation\nvecd[4] = est.mle1(roll)$estdim # MLE with Poisson process\nvecd[5] = est.twonn(roll)$estdim # minimal neighborhood information\n\n# let's visualize\nplot(1:5, vecd, type=\"b\", ylim=c(1.5,2.5), \n main=\"true dimension is d=2\",\n xaxt=\"n\",xlab=\"\",ylab=\"estimated dimension\")\nxtick = seq(1,5,by=1)\naxis(side=1, at=xtick, labels = FALSE)\ntext(x=xtick, par(\"usr\")[3], \n labels = c(\"Ustat\",\"correlation\",\"made\",\"mle1\",\"twonn\"), pos=1, xpd = TRUE)\n```\n\n

\n\nWe can observe that all 5 methods we tested estimated the intrinsic\ndimension around \$d=2\$. It should be noted that the estimated\ndimension may not be integer-valued due to characteristics of each\nmethod.\n\n## Acknowledgements\n\nThe logo icon is made by\n[Freepik](https://www.flaticon.com/authors/freepik/) from\n[www.flaticon.com](https://www.flaticon.com/).The rotating Swiss Roll\nimage is taken from [Dinoj\nSurendran](http://people.cs.uchicago.edu/~dinoj/manifold/swissroll.html)’s\nwebsite.","jobInfo":{"package":"Rdimtools","version":"1.0.4","parsingStatus":"success","parserVersion":1,"parsedAt":"2021-04-18T13:30:05+0000"}}

Description

Usage

Arguments

Examples