pbdML

• Version: 0.1-2
• URL: https://github.com/RBigData/pbdML
• License:
• Author: See section below.

pbdML is an R package containing a collection of machine learning utilities. These functions can be used in serial with native R objects (matrices and vectors) or in parallel with distributed matrices from the pbdDMAT package. Here, we focus on ease of coding and understanding rather than performance (i.e., all code is written in R), and as such this package should primarily be thought of as a demonstration of the capabilities of the pbdDMAT package.

Usage

Functions have the same dispatch whether working with a regular R matrix or with a ddmatrix from the pbdDMAT package.

So for example, to compute a randomized PCA, we can run

rpca(x)

where x is either a matrix or a ddmatrix.

Installation

pbdML requires:

• R version 3.0.0 or higher
• A system installation of MPI
• The pbdMPI and pbdDMAT packages, as well as their dependencies.

Development Version

remotes::install_github("RBigData/pbdML")

Authors

pbdML is authored and maintained by members of the pbdR core team:

• Drew Schmidt
• George Ostrouchov
• Wei-Chen Chen