Enhancements to Revolution R Enterprise's RevoScaleR
R HTML
Switch branches/tags
Nothing to show
Permalink
Failed to load latest commit information.
R Merge pull request #24 from andrie/master May 19, 2014
inst Created additional unit tests Nov 22, 2013
man Merge pull request #24 from andrie/master May 19, 2014
vignettes Added tree visualization vignette. Bumped version to 0.0-05 Sep 24, 2013
.Rbuildignore
.Rhistory Created unit test for rxLinPredError, altered rxSample2Df so that it … Aug 30, 2013
.gitignore
DESCRIPTION Rebuilt package, updated NEWS, bumped version to 0.0-09 May 19, 2014
NAMESPACE Added rxPredictTransform() and isLogFormula() May 19, 2014
NEWS
README.md Update README.md Sep 24, 2013
RevoEnhancements.Rproj

README.md

RevoEnhancements is an R package that adds functionality and utility functions to RevoScaleR. All functions require, and are enhancements to, Revolution R Enterprise.

Additional functionality

Base R Compatability

  • Functions to expand model formulas, e.g. expanding y~. into all dependent variables in the XDF file

Summarising Big Data for easy graphics visualization

  • rxHexBin - hexagonal binning
  • rxBoxPlot

Big Data Mining

  • Discretization and tree discretization
  • Sampling from an XDF into a data frame
  • Calculating test statistics, e.g. accuracy, recall and F1-score

Statistical testing

  • KS statistic
  • t-test

Visualisation

  • Decision trees (plot.rxDTree)
  • Decision forests (plot.rxDTree)

Installation

You can use the devtools package to install RevoEnhancements directly from github.

library(devtools)
install_github("RevoEnhancements", "RevoEnhancements")

RevoEnhancements imports some packages from CRAN, e.g. package hexbin. To install dependencies from CRAN, use:

install.packages("hexbin")

To demonstrate how to use the code, we also use some examples with ggplot2 and mlbench. To install these suggested packages, use:

install.packages(c("ggplot2", "mlbench"))