{"Package":"FTRLProximal","Title":"FTRL Proximal Implementation for Elastic Net Regression","Version":"0.3.0","Authors@R":"person(\"Vilhelm\", \"von Ehrenheim\", email = \"vonehrenheim@gmail.com\", role = c(\"aut\", \"cre\"))","Description":"Implementation of Follow The Regularized Leader (FTRL) Proximal algorithm, proposed by McMahan et al. (2013) , used for online training of large scale regression models using a mixture of L1 and L2 regularization.","URL":"https://github.com/while/FTRLProximal","BugReports":"https://github.com/while/FTRLProximal/issues","Depends":"R (>= 3.2.3)","Imports":"Matrix, methods","License":"MIT + file LICENSE","Encoding":"UTF-8","LazyData":"true","RoxygenNote":"5.0.1","Suggests":"testthat, covr, mlbench","NeedsCompilation":"yes","Packaged":"2017-01-03 07:39:26 UTC; while","Author":"Vilhelm von Ehrenheim [aut, cre]","Maintainer":"Vilhelm von Ehrenheim ","Repository":"CRAN","Date/Publication":"2017-01-03 11:25:34","readme":"# FTRL Proximal\n[![Build Status](https://travis-ci.org/while/FTRLProximal.svg?branch=master)](https://travis-ci.org/while/FTRLProximal)\n[![codecov](https://codecov.io/github/while/FTRLProximal/branch/master/graphs/badge.svg)](https://codecov.io/github/while/FTRLProximal)\n[![CRAN_Status_Badge](http://www.r-pkg.org/badges/version/FTRLProximal)](https://CRAN.R-project.org/package=FTRLProximal)\n[![CRAN Downloads](https://cranlogs.r-pkg.org/badges/FTRLProximal)](https://CRAN.R-project.org/package=FTRLProximal)\n\nThis is an R package of the FTRL Proximal algorithm for online learning of elastic net logistic regression models.\n\nFor more info on the algorithm please see [Ad Click Prediction: a View from the Trenches](https://www.eecs.tufts.edu/~dsculley/papers/ad-click-prediction.pdf) by McMahan et al. (2013).\n\n## Installation\nEasiest way to install is from within `R` using the latest CRAN version:\n\n```r\ninstall.packages(\"FTRLProximal\")\n```\n\nIf you want the latest build from git you can install it directly from github using `devtools`:\n\n```r\ndevtools::install_github(\"while/FTRLProximal\")\n```\n\n## Usage\nSimplest use case is to use the model similar to normal glm with a model formula. \n\n```r\n# Set up dataset\np <- mlbench::mlbench.2dnormals(100,2)\ndat <- as.data.frame(p)\n\n# Train model\nmdl <- ftrlprox(classes ~ ., dat, lambda = 1e-2, alpha = 1, a = 0.3)\n\n# Print resulting coeffs\nprint(mdl)\n```\n\nIt is also possible to update the trained model object once it is trained.\n\n```r\n# Set up first dataset\np <- mlbench.2dnormals(100,2)\ndat <- as.data.frame(p)\n\n# Convert data.frame to model.matrix\nX <- model.matrix(classes ~ ., dat)\n\n# Train on first dataset\nmdl <- ftrlprox(X, dat$classes, lambda = 1e-2, alpha = 1, a = 0.3)\n\n# Generate more of the same data after the first training session\np <- mlbench.2dnormals(100,2)\ndat <- as.data.frame(p)\n\n# Update model using the new data.\nmdl <- update(mdl, X, dat$classes)\n```\n\nFor more example please see the documentation.\n\n## Changelog\n### 0.2\n* Changed from using explicit `lambda1` and `lambda2` parameters to using `lambda` and mixing parameter `alpha`.","repoType":"cran"}

Description

Usage

Arguments

Value

Details